Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
A running list of language access resources on the web
Visit Spanish Language Style Guides and Glossaries on Digital.gov to find more resources for improving how the government communicates with the public in Spanish.
Find more information about how the U.S. Department of Labor approaches plain language for UI on their Plain Language resources page.
To see more plain language definitions of UI terms produced by the Office of UI Modernization at the U.S. Department of Labor, please visit The UI Lexicon.
Learn more about Federal Plain Language Guidelines on plainlanguage.gov
See the Federal Plain Language Guidelines by PLAIN for more examples of how to use active voice effectively.
Visit this American Psychological Association (APA) Style Guide to learn more about Bias-Free Language.
You can learn more about gender-inclusive communication in Spanish from (Re)Nombrar, created by the Ministry of Women, Gender and Diversity of Argentina.
Academic Spanish dictionary: Real Academia Española.
Grijelmo, A., & Merino, J. M. (2019). “More than 555,000,000 of us can read this book without translation”: Mas de 555 millones podemos leer este libro sin traducción = More than 555,000,000 of us can read this book without translation. Taurus. [This collection of essays portrays the richness of the Spanish language and presents a mix of historical, social, cultural, and economic perspectives.]
If you want to design better UI claimant communications, see this Unemployment Insurance Email Template Kit, created by the New Jersey Office of Innovation and Semicolon Design Group.
Gonzales, A.L. (2022, July 7). Why multilingual research matters. Code for America. https://codeforamerica.org/news/why-multilingual-research-matters/
For web accessibility guidance visit World Wide Web Consortium (W3C).
See the W3C's guidelines on “non-text content” and tutorial on optimizing images for the web.
Equitable Name Test: An open source list of names and considerations for names to make systems more equitable.
Promising Practices in State Unemployment Insurance Digital Service Delivery, Beeck Center Digital Services Network, January 2024
New Jersey's Worker-Centered Approach to Improving the Administration of Unemployment Insurance, September 2023
Three years after the pandemic recession, signing up for unemployment still isn’t easy, NPR Marketplace, June 26, 2023
As part of U.S. Digital Response’s pro-bono work to expand government’s capacity to deliver digital services, we offer language access services for state unemployment insurance agencies.
We can help optimize your unemployment insurance system through technical assistance with:
Translations grounded in plain language best practices and policy compliance to improve claimant comprehension
User research with workers in your state to inform data-driven improvements
Consultation on translation tech and workflow to scale multilingual support
These services aim to:
Reduce call center volume by preventing confusion for workers who speak a primary language other than English.
Lower administrative costs by eliminating errors and adjudications.
Increase trust and integrity by improving user experience across languages.
Contact us if you’re interested in language access support for your state.
Email feedback and suggestions for improving our language access resources to:
Originally compiled by Alyssa Levitz with help from Waldo Jaquith
Many of the problems facing Unemployment Insurance agencies today are not unique. Each entity might have a different name, provide different benefit amounts, and require different things from its constituents to access those benefits -- but they can still learn a lot from each other's solutions.
US Digital Response's Unemployment Insurance team works with specific UI agencies on projects and investigations as requested (get in touch with us via our contact form!), and whenever possible, we share our learnings with approval from those partners.
We are also tracking the changes that UI agencies are making on their own or with the help of other organizations. We track these changes based on publicly available information (e.g., press releases, news articles, and website inspection with standard browser developer tools). We are sharing this tracking publicly in the hopes that everyone can benefit from knowing what others are doing.
We have also conducted deep dives on some of these topics:
Identity Proofing (updated: 3/11/21)
Chatbot (updated: 12/22/20)
Appt/call scheduler (updated: 1/21/21)
Online repayments (updated: 11/25/20)
Alabama
SSA cross-matching (2004 Press Release)
No
Yes - opens for the following week on Friday at 5, assigns you a time window for in-person appts
Alaska
None found (SSA presumed)
Yes - IBM's Watson Assistant
Arizona
ID.me implemented in Oct. 2020 (Press Release)
Yes - MSFT BotFramework/WebChat
Arkansas
- None found (SSA presumed) - Unspecified algorithm added to their system in October 2020 (News article)
No
California
ID.me implemented in Sep. 2020 - Results from 1st week - Results from first 3 weeks
Yes - Astute Chat Bot
Colorado
- Experian implemented at some point (News article)
- ID.me implemented Jan. 2021 (article)
Yes - Google Chat Bot - 90% of Qs answered
Yes (youcanbook.me), bottom of the page News article
Connecticut
None found (DMV presumed)
Yes, toward the bottom of the page -- search for "Schedule a Callback"
Jan. 2021, News article
Delaware
None found (DMV presumed)
Yes - ServiceNow
Florida
No - saw an implementation with Google Chat Bot on 11/5/20 but it wasn't there on 12/14/20
Georgia
ID.me implemented in Aug. 2020 (article) - Initial results
Yes - MSFT
Hawaii
None found (DMV presumed)
No
Idaho
- DMV
- ID.me implemented Jan. 2021 (article)
No
Illinois
None found (DMV presumed)
Yes - Google Chat Bot. Some results discussed here.
Indiana
ID.me implemented Fall 2020 (Press release) - Contract link
Yes - In footer of all in.gov pages: faq and live chat through Zendesk - On UI login page: search function with Jira Service Management
Iowa
Using LexisNexis Risk Solutions (web site)
No
Kansas
SSA, backup by DMV - FAQ site
Yes - not sure of vendor (same as MD)
Kentucky
None found (DMV presumed)
Yes - Salesforce
Yes, via Telegov, an NIC Platform (new appts added daily)
Louisiana
ID.me May '21 (article)
No
Maine
ID.me
Yes - Google Chat bot, sometime between mid June and October
Maryland
None found (SSA presumed)
Yes - not sure
Massachusetts
ID.me
Yes - not sure
Michigan
New system of some sort implemented in Jun. 2020 (Press release)
Yes - MSFT BotFramework/WebChat
Phone appt scheduler (10/21 launch) -- site here -- Acuity Scheduling
Minnesota
None found (DMV presumed)
No
Mississippi
ID.me
No
Missouri
ID.me
Montana
- Jun. 2020 added additional mesasures (press release) - Nov. 2020 implemented ID.me (Post)
No
Nebraska
None found (DMV presumed)
No
Nevada
Aug. 2020 began pilot program with ID.me (Article) - Contract - Sept. update
No
New Hampshire
None found (SSA presumed)
No
New Jersey
Yes - Google Chat Bot
New Mexico
None found (SSA presumed)
Yes - Salesforce
New York
ID.me
Yes - Astute Bot
North Carolina
Using multiple sources (article): - NC Data Analytics Center - NASWA Integrity Hub - National Directory of New Hires Wage Crossmatch
-ID.me implemented late 2020/early 2021 (ID.me help article)
Yes - MedChat. (Live chat during business hours, Q&A outside business hours)
North Dakota
- SSA
- ID.me implemented Dec. 2020 (press release)
Yes - MSFT BotFramework/WebChat
Ohio
Oklahoma
Idemia, Dec. 2020 (Press Release)
Yes - MSFT BotFramework/WebChat
No
Oregon
ID.me
Pennsylvania
ID.me implemented in Oct. 2020 (Press Release)
Yes - Genesys
No
Rhode Island
Yes - Twilio
No
South Carolina
ID.me
Yes - MSFT BotFramework/WebChat
no
South Dakota
Yes - MSFT BotFramework/WebChat
No
Tennessee
Yes - Zopim (now Zendesk)
No
Texas
ID.me implemented in Nov. 2020 (Press release)
Yes - AWS & Accenture
No
Utah
Yes - Live chat during business hours - uknown through whom
No
Vermont
Yes - Genesys - Live chat during business hours
No
Virginia
No
No
no
Washington
ID.me implemented Dec. 2020 or Jan. 2021 (website)
Yes - MSFT BotFramework/WebChat
No
West Virginia
Yes - Twilio
No
Wisconsin
Google Cloud implemented in Oct. 2020 (Press Release, Article)
Yes - MSFT BotFramework/WebChat
No
Wyoming
No
No
Guam
No
No
Washington DC
Yes - Google Chat Bot
No
Last updated: September 29, 2023
This Spanish translation guide provides practical guidelines and resources to help public servants improve Spanish translations for unemployment insurance (UI) content. Created through a partnership between U.S. Digital Response (USDR) and the New Jersey Department of Labor (NJDOL), this guide offers strategies and best practices for writing clear, inclusive Spanish tailored to the needs of UI claimants. Our goal is to increase accessibility and build trust with the over 20 million workers in the U.S. whose primary language is Spanish.
This guide was created to help workforce agency administrators and staff produce high-quality Spanish translations for all forms of communication with Spanish-speaking UI claimants. While the guide focuses primarily on digital content, the tips also apply to print materials and verbal communication.
This guide is for anyone involved in or interested in improving the user experience for Spanish-speaking claimants through plain language communication. You might use this guide if you are an adjudicator, call center agent, communications administrator, compliance officer, content administrator, designer, developer, educator, legal aid provider, policy administrator, researcher, social worker, translator, or in a role focused on accessibility.
Unemployment insurance systems are often difficult to navigate, with content written in legalese that is difficult for the average claimant to understand. For claimants whose primary language is not English, this challenge can feel insurmountable without proper translation support.
The need for improving the quality of Spanish translations is clear when looking at key challenges:
The United States has over 42 million Spanish speakers from diverse regions, each with distinct vocabularies. Most existing Spanish translations do not reflect this diversity and may exclude many Spanish speakers.
Translations of public-facing government services are often not plain or simple.
Translations tend to have errors, either from a lack of native speakers or reliance on unchecked machine translation.
Unclear wording can lead to people providing incomplete or incorrect information. This results in delayed benefits, denials, or even repayment directives for benefits they qualify for.
Poor translations decrease user satisfaction and trust while increasing disparities in benefits access.
This guide was created using an inclusive, human-centered design process. We focused on uplifting and learning from Spanish-speaking communities by conducting usability testing with a diverse group of native Spanish speakers to develop translations for a new UI claimant intake form. The insights from this research, along with expertise from the NJDOL's bilingual call center team, directly informed the recommendations in this guide.
Our goal is to improve UI communications and better serve Spanish-speaking workers. In pursuing this goal, we grappled with some key questions: What happens when plain English is translated into another language? How do we maintain a human-centered approach and achieve human-centered results across different cultures and vocabularies? How is equity expressed?
Translation inherently involves conversion—a complex shift. This approach views translation not just as word conversion, but as an opportunity to enrich our understanding of claimants’ lived experiences. By putting inclusive design into action and centering Spanish-speaking users, we aim to increase access, build trust, and advance equity in public services. The strategies in this guide provide ways to effectively translate plain English into plain Spanish, converting “the plain” to “lo simple.”
We’d love to hear from you! Send feedback and suggestions to:
"Plain language makes it easier for the public to read, understand, and use government communications." plainlanguage.gov
This guidance follows Federal Plain Language Guidelines, with additional considerations for Spanish grammatical styles and cultural norms.
In general, plain language writing should:
Use simple, common words and sentences
Avoid unnecessary words and jargon
Present information in a logical order
Focus on the audience's needs
Have a conversational tone
Adhere to web accessibility standards
Consider what devices people may be using to view the content
When writing UI content, consider the target audience's vocabulary, education level, and situation. The goal is not to write in elementary terms, but rather use language appropriate for the readers.
In addition to the principles listed above, UI content should aid comprehension and guide users to take the desired action. To do this effectively it should::
Aim for an 8th-grade reading level
Maintain consistent terminology
Favor short, clear sentences
Use informative headers to structure content
Explain concepts directly to readers
Make text scannable with bullet points, bolding, etc.
Use active voice and action verbs for calls to action
Maintain an objective, yet empathetic tone
Learn more about the U.S. Department of Labor’s recommended approach to plain language for unemployment insurance on their Plain Language resources page.
The following sections are organized from simple to complex: First, we focus on words — the building blocks of communication. Then we move on to sentences and paragraphs, followed by guidance on capitalization and punctuation at the end.
When writing content, follow these guidelines for clear, concise sentence structure:
Be explicit. Avoid ambiguous phrases that could have multiple meanings.
Use short, direct sentences.
Frame sentences from a positive standpoint - let users know what you want them to do, not what you don’t want them to do.
When translating to Spanish:
Aim to match the English content in length and organization. The versions won't be identical but should be similar.
Spanish is the official language in 21 countries, primarily spoken in South and Central America and Spain, with significant use also in parts of Africa and Asia. With over 577 million people worldwide speaking Spanish as their first or second language, it is no surprise that the language varies across regions.
In the United States, nearly 18% of the population speaks Spanish, and this segment is rapidly growing. Even though Spanish speakers can understand each other across dialects, words and terms often differ. Conducting user testing helps determine which terms will be most universally understood.
Providing synonyms and alternative vocabulary ensures content resonates with users from different countries and age groups. Below are some words for which we suggest including synonyms in Spanish, based on user testing—for some of those cases we recommend including the word in English.
Spanish pronouns, nouns, and adjectives are gendered as female or male, usually ending with “A” for female and with “O” or “E” for male.
Gendered grammar can reinforce biases. If we say “doctores” (male) and “niñeras” (female), we are reinforcing the assumption that, by default, all doctors are men and all nannies are women.
Today there are new ways to be more inclusive in Spanish, such as saying “les niñes” instead of “los niños.” However, this particular usage is not yet widespread. Until this approach is more widely adopted, we do not recommend using this format in government communications with Spanish speakers.
We recommend the following approaches, in order of preference:
Labeling someone by a disease or disability can feel dehumanizing. Use language that upholds dignity.
When discussing disability in both in English and Spanish:
Focus on the person first, not the disability.
Say "a person with a disability" rather than "disabled person."
In Spanish, say “tiene una discapacidad” (has a disability) rather than “es discapacitado/a” (is disabled).
Use impartial, non-judgmental language.
Indicate that someone “has multiple sclerosis”, not that they “suffer from multiple sclerosis.”
Use terms like “accessible parking” instead of “handicapped parking” and “accessible restroom” instead of “disabled restroom.”
Specify a disability only if relevant. Avoid generalizations.
Avoid stigmatizing terms related to disability or mental health (e.g. crazy, dumb, lame, insane, or stupid) or sensory issues (e.g. blind spot or tone deaf).
Ask individuals and groups directly about preferred terminology, as language evolves.
Use "incapacidad laboral" when discussing workers' inability to work as a result of a physical or mental illness.
The distinction matters. "Incapacidad laboral" focuses on impacts in a work context. Using the right term avoids incorrect assumptions that a disability inherently precludes working.
Refer to "people" or "the public" when discussing government services.
Be specific about groups without stereotyping (even stereotypes that may appear to be positive). Avoid imagery or language reinforcing stereotypes. For example:
Use “people who need healthcare” or “people who need to access government services online”.
Avoid the term nonwhite, or other terms that treat whiteness as a default.
Use adjectives, not nouns, for race or ethnicity. For example, a “Hispanic person” not “a Hispanic.”
Instead, we use positive framing, such as "primary Spanish speaker" or "people whose primary language is other than English" (PLOE)—a term coined by language access researcher Dani Carillo.
Avoid idioms and figurative phrases that may not translate literally into Spanish. Idiomatic expressions like "it's raining cats and dogs" can seem confusing or nonsensical when translated word-for-word.
Be clear and specific. Avoid ambiguities or phrases that can be interpreted in multiple ways.
Use common words your audience will understand. Specialized vocabulary excludes many readers. Know your audience, and identify where you can replace specialized terms with simpler words.
If specialized vocabulary must be used for compliance or other reasons, define it using common words when possible:
Provide definitions where they are critical for readers to understand the action to take next.
Use definitions a non-expert would understand.
Provide a glossary of terms that is easy for users to find, as UI systems involve many specialized terms.
Use helper text to educate users when clarification is needed. Provide contextual copy to explain concepts like legal terms, UI jargon (e.g. "able & available"), or words that don't directly translate to Spanish.
For example:
Suffixes like "Jr." are common in English but not Spanish names. Explain that users should enter their name, including any suffixes, as they appear on their government ID.
For U.S.-specific programs like Paid Family and Medical Leave, provide explanatory details since these programs will likely be unfamiliar to many users, especially those who are new to the country. Assume no prior knowledge and educate through helper text.
In Spanish, there are three ways to say "you" - “usted”, "tú", "vos." This can be confusing for English speakers. The right pronoun to use depends on the level of formality and the region.
Usted
Use "usted" when you want to be polite and formal. It works for everyone. You can't go wrong with "usted"!
In writing, "usted" is the most common "you." But consider your reader - "tú" or "vos" may be better for young people.
To say the plural "you," use "ustedes" for all people.
Tú and Vos
“Tú” and “vos” are more casual
“Tú” is common in Colombia, Cuba, Ecuador, Panama, Peru, and Venezuela. Its plurals are “vosotras” (female) and “vosotros” (male).
“Vos” is more common in Argentina, Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua, and Paraguay. Its plural is “ustedes.”
In summary, there is no wrong choice! Ask what the person prefers, or consider the relevant characteristics of your audience. When in doubt, use "usted."
Use active voice and avoid third person when possible. This makes it clear who needs to take action. Passive voice and third person create distance and ambiguity.
When translating content from English to Spanish, leverage cognates to improve comprehension. Cognates are words that look and sound similar in both languages while sharing the same meaning. For instance, words like "family" and "familia" have the same Latin roots. Cognate words also tend to have similar length and cadence in both languages.
When translating UI content, writers should identify and use cognate words where appropriate. Using cognate pairs helps keep the English and Spanish versions aligned more closely. This aids comprehension and retention for bilingual English and Spanish-speaking audiences.
In some cases, including the English word alongside the Spanish translation can aid comprehension. This is helpful when:
Users may only know certain keywords in English if used primarily in English-dominant contexts.
Bilingual users in the U.S. may not recognize some specialized terms in Spanish (e.g. "indemnización" meaning "severance").
To avoid confusion:
Use "Despedido/a con causa" to mean "fired". This conveys "dismissed with cause."
Use "Despedido/a sin causa" for "laid off". This means "dismissed without cause."
Include the English word alongside each Spanish phrase for extra clarity
There are also some words that are best left in English rather than using Spanish translations, especially for those living in the U.S.
Examples to leave in English:
Call center
Full-time
Part-time
Online
The name of the states of the United States. For example “New Jersey” instead of “Nueva Jersey”
Using fewer and shorter words makes text easier to read. It also aids translation, as Spanish content tends to be longer. Here are some tips:
Use descriptive titles that clearly communicate the content that follows. This helps orient the user and reduces cognitive load for people who may already feel stressed when seeking UI benefits.
Recommendations:
Keep titles short yet descriptive. Overly long titles can be confusing and difficult to read on small screens or mobile devices.
Add explanatory subtitles to break up long sections of content. Subtitles can split lengthy forms into smaller chunks of fields or questions, or segment dense blocks of text on web pages into more digestible components. This makes it easier for users to focus on each piece of content.
Acronyms work slightly differently in Spanish compared to English. In English, once an acronym is defined, you can use the acronym by itself going forward. However, in Spanish, if the acronym originated from English, best practice is to include the full Spanish translation plus the English acronym each time, followed by "por sus siglas en inglés" to indicate it's an English acronym.
Capitalization rules differ between English and Spanish. In English, most words are capitalized in titles, proper nouns, etc. However, Spanish uses capital letters more sparingly.
In Spanish, capitalization is required in certain cases like:
However, Spanish does not require capital letters in some cases where English does:
Punctuation rules are generally similar between English and Spanish, with some key differences to note.
Exclamation points: Spanish requires the inverted exclamation point ¡ at the beginning of a sentence, versus only at the end in English.
Question marks: Spanish requires the inverted question mark ¿ at the beginning of an interrogative sentence, versus only at the end in English.
Oxford commas: The comma before "and" or "or" in a list is typically not used in Spanish like it is in English (although debatable - search “Oxford comma controversy” for a fun rabbit hole)
Periods/commas with quotes: In English, punctuation goes inside the quotation marks, while in Spanish it always goes outside the closing quotation mark. e.g.: “El formulario mejoró mucho”, dijo un participante.
Complex | Simple |
---|
Negative | Positive |
---|
Word in English | Words in Spanish |
---|
See our comprehensive list of recommended synonyms in our .
English | Spanish |
---|
Gender-specific (don’t say) | Gender inclusive (say) |
---|
Learn more about gender-inclusive communication in Spanish from , created by the Ministry of Women, Gender and Diversity of Argentina.
Don’t say | Say |
---|
A common mistake in Spanish is to use the words “incapacidad” and “discapacidad” interchangeably when these words have different meanings. "Incapacidad" translates to an inability or lack of capability. "Discapacidad" translates to a disability.
This table shows the definitions according to .
English | Spanish |
---|
Don’t say | Say |
---|
Don’t say | Say |
---|
You may have noticed we avoid phrases that frame people who need language access as deficient. For example, we avoid terms like "limited English proficiency" or "non-English speaker." Similarly, "English as a second language (ESL)" has negative connotations, especially for young people in the U.S. school system.
Visit this American Psychological Association (APA) Style Guide to learn more about .
Don’t say | Say |
---|
Specialized Vocabulary | Common vocabulary |
---|
Specialized term | Definition |
---|
Visit to find a full list of plain language definitions for common unemployment insurance terms, produced by the Office of UI Modernization at the U.S. Department of Labor.
Topic | Helper text examples |
---|
Don’t say | Say |
---|
See the for more examples of how to use active voice effectively.
Cognates in English | Cognates in Spanish |
---|
Watch out for "false cognates" - words that sound similar but have different meanings in English and Spanish.
False cognates in English | False cognates in Spanish |
---|
The Spanish word “despedido/a"can be confusing as it can be translated to either "fired" or "laid off."
English | Spanish |
---|
Don’t say | Say |
---|
Non-descriptive title | Descriptive title |
---|
English | Spanish |
---|
Capitalization is required for | Examples |
---|
Capitalization is not required for | Examples |
---|
English | Employment opportunities | Job offers |
Spanish | Oportunidades laborales | Ofertas de trabajo |
English | Don’t submit your form if you don’t know your exact address. | Check that your address is correct before submitting your form |
Spanish | No envíe su formulario si no sabe su dirección exacta. | Verifique que su dirección sea correcta antes de enviar su formulario. |
Check your pay stubs or W2 to confirm your employer's name. | Revise sus talonarios de pago/recibos (pay stubs) o su W2 para confirmar el nombre de su empleador/a. |
Complete your union membership information. | Complete los datos de su gremio/unión de trabajadores/as. |
“Worker” (female) | La trabajadora |
“Worker” (male) | El trabajador |
“Employer” (female) | La empleadora |
“Employer” (male) | El empleador |
Refers to groups of people instead of particular individuals, and uses more nouns instead of gendered adjectives | Los trabajadores | El grupo de trabajadores |
Los hispanos | Personas hispanas |
Dueño | Persona propietaria |
Manténgase contactado con su seguro médico | Manténgase en contacto con su seguro médico |
¿Es ciudadano de los Estados Unidos? | ¿Tiene la ciudadanía de los Estados Unidos? |
Los directores decidieron… Los secretarios harán… Los presidentes son.. | Dirección decidió… Secretaría hará… Presidencia se compone por.. |
Use “usted” | El solicitante debe completar… | Usted debe completar… |
Use non personal forms | El solicitante debe completar… | Es necesario completar… |
Say/write both | Los trabajadores | Las trabajadoras y los trabajadores |
Combine feminine and masculine (or masculine and feminine), using a “/,” when any of the previous options are not possible | Hermano | Hermana/o |
Despedido | Despedida/o |
Empleador | Empleador/a |
Doctor | Doctor/a |
Niñeras | Niñeros/as |
English | People who suffer from multiple sclerosis | People who have multiple sclerosis |
Spanish | Personas que sufren de esclerosis múltiple | Personas que tienen esclerosis múltiple |
English | Insane | Unreasonable, improper |
Spanish | Demente | Poco reasonable, inapropiado |
Discapacidad | A situation where a person who, due to their physical, sensory, intellectual or mental conditions, encounters challenges in their participation and social inclusion. | Situación de la persona que, por sus condiciones físicas, sensoriales, intelectuales o mentales duraderas, encuentra dificultades para su participación e inclusión social. |
Incapacidad | Lack of ability to do, receive or learn something. | Falta de capacidad para hacer, recibir o aprender algo. |
Incapacidad Laboral | A situation of illness or physical or mental illness that prevents a person, temporarily or permanently, from carrying out a professional activity and which normally entitles them to a social security benefit. | Situación de enfermedad o de padecimiento físico o psíquico que impide a una persona, de manera transitoria o definitiva, realizar una actividad profesional y que normalmente da derecho a una prestación de la seguridad social. |
English | Citizen, resident | People, the public, users, folks, the audience, the community |
Spanish | Ciudadano, residente | La gente, el público, las y los usuarios, la audiencia |
English | Colombians | People from Colombia |
Spanish | Colombianos | Personas de Colombia |
English | Enter your number | Enter your phone number |
Spanish | Ingrese su número | Ingrese su número de teléfono |
English | If the best option is to use jargon because any other term risks losing the precise meaning, explain the concepts the first time the word is used. | If the best option is to use specialized vocabulary because any other term risks losing the precise meaning, explain the concepts the first time the word is used. |
Spanish | Si la mejor opción es utilizar la jerga porque cualquier otro término perdería precisión en su significado, explique los conceptos la primera vez que los use. | Si la mejor opción es usar vocabulario especializado porque cualquier otro término perdería precisión en su significado, explique los conceptos la primera vez que los use. |
English | I may receive sanctions if benefits are improperly paid as a result of providing fraudulent information. | I may be fined if benefits are improperly paid if I provide false information. |
Spanish | Puedo recibir puniciones si los beneficios se pagan indebidamente como resultado de proporcionar información fraudulenta. | Puedo recibir multas si los beneficios se pagan indebidamente si doy información falsa. |
English | Maximum benefit amount* | The total amount of benefits you may get during a benefit year. |
Spanish | Cantidad máxima del beneficio | Translation TBD pending usability testing |
English | Disability | For Unemployment Insurance, "disabled person" is defined by any of the following cases:
|
Spanish | Incapacidad | Para el Seguro de Desempleo, “persona incapacitada” se define por alguno de los siguientes casos:
|
English | Paid Family and Medical Leave Program | The State of XXXX Paid Family and Medical Leave Program could cover one or more of these cases:
|
Spanish | Programa de licencia para familias | El programa de licencia para familias del Estado de XXXX podría cubrir uno o más de estos casos:
|
English | The form must be completed. | Complete the form. |
Spanish | El formulario debe ser completado. | Complete el formulario. |
English | Claims must be submitted before the end of the month. | Submit your claim before the end of the month. |
Spanish | Los reclamos deben ser enviados antes de fin de mes. | Envíe su reclamo antes de fin de mes. |
Application process | Proceso de aplicación |
Calculate | Calcular |
Employment | Empleo |
Form | Formulario |
Organization | Organización |
Terminate | Terminar |
Actual | Actual (means “current”) |
Contest | Contestar (means “to answer”) |
Exit | Éxito (means “success”) |
Lecture | Lectura (means “reading”) |
Parent | Pariente (means “relative”) |
Why were you fired, discharged, or suspended? | ¿Por qué fue despedido/a con causa (fired) o suspendido/a? |
Why were you laid off? | ¿Por qué fue despedido/a sin causa (laid off)? |
Severance pay is payment you receive due to policy, contract, or other note about your job ending. | Los pagos por despido (severance pay) son pagos que usted recibe por una póliza, contrato, o similar, al terminar su trabajo. |
Use shorter synonyms | English | Utilize | Use |
Spanish | Utilizar | Usar |
Limit adjectives
| English | Is able to | Can |
Spanish | Es capaz de | Puede |
English | A significant number of… | Several, many, enough |
Spanish | Número important de… | Varios, pocos, bastantes |
Avoid repetitions
| English | Bring your passport so we can check the visa in your passport. | Bring your passport so we can check your visa. |
Spanish | Traiga su pasaporte para verificar la visa en su pasaporte | Traiga su pasaporte para verificar su visa. |
English | The X Agency and the Y Agency worked together on a joint project to improve… | The X and Y Agencies worked together on a project to improve… |
Spanish | La Agencia X y la Agencia Y trabajaron juntas en un proyecto conjunto para mejorar… | Las Agencias X e Y trabajaron juntas en un proyecto para mejorar… |
Avoid unnecessary words | English | These sections describe types of information that would satisfy the application requirements for this grant program. | These sections tell you how to meet the requirements for this grant program. |
Spanish | Estas secciones describen tipos de información que cumplirían con los requisitos de solicitud para este programa de subvenciones. | Estas secciones le indican cómo cumplir con los requisitos para este programa de subsidio. |
English | Plain Language | Writing Plain Language |
Spanish | Lenguaje Simple | Escribiendo en Lenguaje Simple |
English | Application Form | Unemployment Insurance Application Form |
Spanish | Formulario de solicitud | Formulario de solicitud del Seguro de Desempleo |
Centers for Disease Control and Prevention (CDC) | Centros para el Control y la Prevención de Enfermedades (CDC, por sus siglas en inglés) |
This guide follows “Federal Plain Language Guidelines” by the Plain Language Action and Information Network (PLAIN). | Esta guía se adhiere a las “Pautas Federales de Lenguaje Simple” de la Red de Información y Acción en Lenguaje Simple (PLAIN, por sus siglas en inglés). |
The first word of titles and sentences | Manténgase en contacto con su seguro médico |
People’s first and last names | Fernanda Williams, Greg Brown |
Geographical places like countries, cities, continents, planets | Estados Unidos, California, Brooklyn, Asia, Júpiter |
Holidays | Año Nuevo, Navidad |
Acronyms | USCIS |
Institutions
| Ministerio de Diversidad y Género, Museo Nacional de Arte |
Languages | español, chino, inglés, |
Days of the week and months | lunes, martes, miércoles enero, febrero, marzo |
Nationalities | latina/o americana/o, estadounidense |
Religions | budismo, catolicismo, judaísmo |
This style guide is a living document, a work in progress. We welcome feedback on how to make it better!
If you’d like to see this resource made available in additional languages, let us know how we can help.
Language is intrinsically complex, continuously evolving, and intertwined with human identity. We took on this project earnestly and enthusiastically, grateful for a chance to participate in language's ongoing evolution. Aware that word choice shapes how readers understand complex ideas and instructions, potentially expanding or limiting access to resources like unemployment benefits, we approached this work with intention and humility. It is our hope that the care behind this work shines through. Just as a key unlocks a door, we hope these guidelines can unlock connection, dialogue, and trust.
This work was made possible thanks to:
The New Jersey Department of Labor and New Jersey Office of Innovation, who entrusted USDR as a partner in Spanish content development. We especially thank the bilingual call center agents who generously shared their wisdom to provide a solid foundation.
Barbara Niveyro, USDR Content Strategist, for her curiosity, boldness, and native Spanglish writing.
UX researchers Holly Worthen, Joe Wiltberger, and Leyda Vega Hughes for volunteering their time and efforts.
Marcie Chin, USDR Product Lead for Unemployment Insurance, for her leadership and diligence.
Madel Leal, USDR Lead Language Access UX Researcher, for her advice and insights.
We are eager to continue co-creating with UI administrators and workers to broaden language access and help government keep up with the critical needs of the public - now and in the future.
Thank you!
¡Muchas gracias!
Visit Spanish Language Style Guides and Glossaries on Digital.gov to find more resources for improving how the government communicates with the public in Spanish.
Find more information about how the U.S. Department of Labor approaches plain language for UI on their Plain Language resources page.
To see more plain language definitions of UI terms produced by the Office of UI Modernization at the U.S. Department of Labor, please visit The UI Lexicon.
Learn more about Federal Plain Language Guidelines on plainlanguage.gov
See the Federal Plain Language Guidelines by PLAIN for more examples of how to use active voice effectively.
Visit this American Psychological Association (APA) Style Guide to learn more about Bias-Free Language.
You can learn more about gender-inclusive communication in Spanish from (Re)Nombrar, created by the Ministry of Women, Gender and Diversity of Argentina.
Academic Spanish dictionary: Real Academia Española.
Grijelmo, A., & Merino, J. M. (2019). “More than 555,000,000 of us can read this book without translation”: Mas de 555 millones podemos leer este libro sin traducción = More than 555,000,000 of us can read this book without translation. Taurus. [This collection of essays portrays the richness of the Spanish language and presents a mix of historical, social, cultural, and economic perspectives.]
If you want to design better UI claimant communications, see this Unemployment Insurance Email Template Kit, created by the New Jersey Office of Innovation and Semicolon Design Group.
Gonzales, A.L. (2022, July 7). Why multilingual research matters. Code for America. https://codeforamerica.org/news/why-multilingual-research-matters/
For web accessibility guidance visit World Wide Web Consortium (W3C).
See the W3C's guidelines on “non-text content” and tutorial on optimizing images for the web.
Equitable Name Test: An open source list of names and considerations for names to make systems more equitable.
“We want our documents to help the audience get information, comply with requirements, and apply for benefits with the minimum possible burden.” - Federal Plain Language Guidelines (PDF Report - page 94)
Optimizing for readability is important for effectively engaging all audiences, especially considering that an estimated 50% or more of users now interact with content on mobile devices. It can also be particularly beneficial for people whose primary language is other than English for a few key reasons:
Cognitive load: Reading in a second language has higher cognitive demand. Well-formatted, scannable content reduces the processing load. Use headers, lists, and other formatting techniques.
Visual aids: Supplementing text with images, charts, diagrams, etc. provides additional context to help users quickly understand what documents they need to provide or what action they need to take. Include relevant photos, graphics, or illustrations.
Consistency: A clear, logical structure and organization help users anticipate content flow. Use a consistent format and organization.
Use visual hierarchy, white space, and lists to draw attention to key information and break up dense text, making it more scannable for Spanish readers.
Highlight information by:
Enclosing text in a box
Changing type size, weight, or font
Using bulleted or numbered lists
Additionally, structure content following the inverted pyramid style:
What happened (what)
Who was involved (who)
Where it took place (where)
When it happened (when)
Why it happened (why)
How it happened (how)
For best practices on designing UI claimant communications, see this Unemployment Insurance Email Template Kit, created by the New Jersey Office of Innovation and Semicolon Design Group.
Images can simplify complex ideas, make content more engaging, and serve as helpful visual aids. By addressing both accessibility and optimization, you can ensure images are effective for all readers.
Provide descriptive alt text to convey the purpose and context of each image. This makes content accessible for those using screen readers or slow connections.
Focus alt text on describing the meaningful context of the image, rather than just decorative details. The World Wide Web Consortium (W3C) guidelines recommend alt text should serve an "equivalent purpose" as the image itself.
Compress image file sizes to optimize for fast loading. This improves the user experience for mobile users or those with slow connections.
For more guidance, see the W3C's guidelines on “non-text content” and tutorial on optimizing images for the web.
Dates can be a source of confusion in English-to-Spanish contexts due to different date formatting conventions. In English, dates are typically written in MM/DD/YYYY format. However, in Spanish, the more common format is DD/MM/YYYY.
To prevent users from entering incorrect dates in form fields, use these UX best practices:
Include clear labels alongside date input fields to specify the desired day, month, and year format. This helps users enter dates accurately.
Display dates in unambiguous ways like spelling out months and adding labels for day, month, year. For example, "August 8, 2022" or “8 August 2022” rather than "08/08/2022".
Use a calendar picker that allows dates to be selected visually rather than typed in. This reduces errors from typing dates in the wrong format.
Where date fields cannot be calendar pickers, accept a wide range of date formats and validate the entry. Provide clear error messaging if the wrong format is used.
Claimant communication, fact-finding content, and form design, created and tested in English and Spanish
View all Retroactive PUA Eligibility materials as Google Documents and download in your preferred format.
U.S. Digital Response worked with a state workforce agency to create and test a fact-finding flow for claimants who had been denied Pandemic Unemployment Assistance (PUA) benefits before three new provisions were introduced by the US DOL in UIPL 16-20, Change 5 in February 2021.
State workforce agencies across the U.S. must reopen PUA applications to claimants who were previously denied benefits and determine if they are eligible to receive retroactive payments.
Maximize claimant response rate
Maximize application success rate
Minimize impact on support staff
To help additional state workforce agencies achieve these goals while saving a significant amount of implementation time and resources, USDR has made the materials we’ve created for our state partner free to use for any state, district, or territory in the U.S.
These materials include:
Claimant communication strategy
Outreach content that can be used for both digital and paper notifications to help claimants understand why they are being contacted and designed to mitigate for the perception of fraud/spam.
Fact-finding content strategy
All 19 eligibility questions, each accompanied by legally compliant helper text optimized to increase readability, comprehension, and accuracy of responses.
Fact-finding form design
Created for states using the FAST UI platform and tested for usability, these prototypes can be used as a blueprint for any state, regardless of platform or vendor.
"This was an opportunity to move away from solely the regulatory side of UI into more of a customer-centric approach. I think it’s going to really establish a new path forward and increase the level of confidence of our stakeholders – hearing from actual users and reading their feedback does that work."
– State workforce agency adjudication manager
By testing these materials with both English and Spanish-speaking claimants, we believe that they will reduce the risk that typically accompanies launching a new service for workers in need of unemployment insurance benefits. In taking a multilingual approach, we are committed to supporting state efforts to improve equitable access to benefits, as primary Spanish speakers comprise the 2nd largest language demographic in the country.
You can:
Hit the "Next" button below to continue browsing our Retroactive PUA Eligibility materials
View all Retroactive PUA Eligibility materials as Google Documents and download in your preferred format
Download the report to learn about our methodology and research insights.
We are also available to consult on implementation or work with your staff on customizing these materials for your state.
Let us know if you’re interested in learning more about how USDR can support your Retroactive PUA Eligibility efforts by submitting a request here.
Outreach content that can be used for both digital and paper notifications to help claimants understand why they are being contacted and designed to mitigate for the perception of fraud/spam.
View all Retroactive PUA Eligibility materials as Google Documents and download in your preferred format.
Subject
Notice: You may be eligible to reapply for Pandemic Unemployment Assistance
Body
Earlier this year, the U.S. Department of Labor expanded eligibility provisions for Pandemic Unemployment Assistance (PUA) to include three new COVID-19 related reasons under which an individual may self-certify:
Individuals who refuse to return to work that is unsafe or to accept an offer of new work that is unsafe
Certain individuals providing services to educational institutions or educational service agencies
Individuals experiencing a reduction of hours, or a temporary or permanent layoff.
These provisions apply retroactively to the beginning of the PUA program, although individuals who did not file an initial PUA claim on or before December 27, 2020 are limited to weeks of unemployment beginning on or after December 6, 2020.
Based on these new eligibility provisions, you could now be eligible for PUA payments. The [INSERT STATE WORKFORCE AGENCY NAME] will notify you via email shortly when you will be able to self-certify on its website.
What this means for you:
You should certify online at [INSERT URL HERE] for a chance to establish eligibility when you hear from us again. Failure to do so could mean not knowing if you were incorrectly denied previously and are owed weeks of PUA funds.
Thanks and we look forward to hearing from you soon!
Subject
Reminder: Deadline for Pandemic Unemployment Assistance self-certification
Body
This email is to remind you that your deadline for establishing PUA eligibility is approaching.
Please log on to [INSERT URL HERE] to respond with your information by [date].
Don't wait. This is your chance to determine if you were incorrectly denied PUA previously and are owed weeks of PUA funds.
Subject
Final Reminder: Deadline for Pandemic Unemployment Assistance self-certification
Body
This is your last chance to establish PUA eligibility in our system before your [date] deadline.
Log on now to [INSERT URL HERE] to respond with your information by [date].
Once your deadline passes, you will no longer be able to determine if you were incorrectly denied PUA previously and are owed weeks of PUA funds.
Form intro content plus all 19 eligibility questions, each accompanied by legally compliant helper text optimized to increase readability, comprehension, and accuracy of responses.
Issue: [INSERT ISSUE NAME]
Slug: PUA Expanded Eligibility (2020)
Title: Background
We are reaching out to you because we recognize you as someone who was previously denied PUA eligibility in our records, but may be now eligible for payments based on expanded eligibility provisions for PUA in the past year.
These provisions apply retroactively to the beginning of the PUA program. As a result, you must respond to the following questions to establish eligibility even if you answered them in the past. You will also recognize previous questions found on weekly claim forms. Certify for all weeks and make sure to answer questions 17 to19 related to the newest eligibility provisions covering:
Individuals who refuse to return to work that is unsafe or to accept an offer of new work that is unsafe;
Certain individuals providing services to educational institutions or educational service agencies;
Individuals experiencing a reduction of hours, or a temporary or permanent layoff.
Note: individuals who did not file an initial PUA claim on or before December 27, 2020 are limited to weeks of unemployment beginning on or after December 6, 2020.
Unless you absolutely need to, please do not call us with questions. Most questions are answered on our website. You should be able to complete this form on your own.
1. Did your place of employment close as a direct result of the COVID-19 public health emergency?
Answer yes if:
your place of employment closed due to the COVID-19 public health emergency
you cannot return to regular hours of work due to the COVID-19 emergency declaration or social distancing protocols
Answer no if:
your place of employment did not close at any time during the COVID-19 public health emergency
your place of employment closed but you can return to regular hours of work this week
your place of business closed for reasons not related to the COVID-19 emergency declaration or social distancing protocols
2. Are you an independent contractor or self-employed individual, and has the COVID-19 public health emergency severely limited your ability to do your customary work activities?
Answer yes if:
you are an independent contractor or self-employed individual AND had a reduction in your usual work to the point you were forced to suspend activities during this week
Answer no if:
you are not an independent contractor or self-employed individual
you are an independent contractor or self-employed individual, and you have been able to work your regular number of hours this week
you are an independent contractor or self-employed individual, and you can perform your customary work activities (business is not significantly slower than normal)
3. Are you unable to work from home or anywhere because you are the main caregiver to a child or household member who needs your constant and ongoing care; and they cannot attend school, daycare or other facility that is closed as a direct result of the COVID-19 public health emergency?
Answer yes if:
you cannot work because you need to provide ongoing and constant care to a child or household
you do not receive paid leave benefits under the Families First Coronavirus Responses Act (FFCRA)
the school, daycare or other facility is closed due to the COVID-19 public health emergency (school is considered closed when it is fully online and does not provide in-person classes, or it offers a mixture of online and in-person classes with students attending some days remotely)
Answer no if:
you can work from home while being the main caregiver to a child or household member
the school, daycare or other facility is open
you chose to allow the child to attend class remotely (school is considered open when it requires students to be physically present or when students have the option to stay home and attend online classes or school in-person full time)
4. Are you unable to reach your place of employment because of a quarantine imposed by a state or municipal order restricting travel as a result of the COVID-19 public health emergency?
Answer yes if:
your county is restricting travel and therefore you are not able to work
Answer no if:
your county is not in a phase where travel is restricted
5. Are you unable to reach your place of work because you have been advised by a healthcare provider to self-quarantine due to concerns related to COVID-19?
Answer yes if:
you were advised to self-quarantine during this week due to COVID-19
Answer no if:
you have not been advised to self-quarantine this week due to COVID-19
6. Were you scheduled to start a job in Washington state but no longer have the job or are unable to reach the job as a direct result of the COVID-19 public health emergency?
Answer yes if:
you were hired and scheduled to start a job, but no longer have or cannot reach that job due to the COVID-19 public health emergency
Answer no if:
you were not hired and scheduled to start a job
you are employed and can reach your job
7. Did you have to quit as a direct result of COVID-19?
Answer yes if:
you quit your job as a direct result of COVID-19
Answer no if:
you did not have to quit your job due to COVID-19
you quit your job due to COVID-19 but can return to work this week
8. Were you diagnosed with COVID-19?
Answer yes if:
you were diagnosed with COVID-19 and are still unable to return to work (a qualifying diagnosis does not require a positive test. Any diagnosis from a qualified medical professional, including one made via phone or telehealth, is sufficient.)
Answer no if:
you were diagnosed with COVID-19 but were able to return to work this week
you were not diagnosed with COVID-19
9. Do you have symptoms of COVID-19 and are seeking a medical diagnosis?
Answer yes if:
you had COVID-19 symptoms, were seeking a medical diagnosis, and were unable to return to work for this reason
Answer no if:
you did not have COVID-19 symptoms
you had COVID-19 symptoms but did not seek a medical diagnosis
you had COVID-19 symptoms, sought a medical diagnosis, and were able to return to work because your diagnosis was negative
10. Has an individual in your household been diagnosed with COVID-19?
Answer yes if:
someone in your household has been diagnosed with COVID-19 and you are unable to return to work this week for this reason
Answer no if:
no one in your household has been diagnosed with COVID-19
11. Are you providing constant and ongoing care for a family member or household member who has been diagnosed with COVID-19?
Answer yes if:
you provide constant and ongoing care for a family or household member who has COVID-19
Answer no if:
you do not provide constant and ongoing care to a family or household member who has COVID-19
12. Did you become the breadwinner or major support for a household because the head of household died as a direct result of COVID-19?
Answer yes if:
you became a major source of income for your household and are unable to find work
Answer no if:
you became a major source of income due and are employed this week
you did not become a major source of income after the head of household died
13. Were you offered the option to telework (work remotely or from home) for your usual number of hours with pay during the COVID-19 public health emergency?
Answer yes if:
you received the option to telework and you can do so this week for your usual number of hours with pay
Answer no if:
you did not receive the option to telework by your usual employer
14. Can you telework (work remotely or from home) this week for your usual number of hours with pay during the COVID-19 public health emergency?
Answer yes if:
you can telework your usual number of hours with pay this week
Answer no if:
you cannot telework your usual number of hours with pay this week
you cannot telework because you are providing constant and ongoing care for your children while schools or daycares are closed
you cannot telework due to the nature of your job (for example, if you work in construction or at a restaurant)
you cannot telework due to domestic violence, sexual violence, or stalking
15. Did you apply for or receive any private income protection insurance or supplemental unemployment benefits?
Answer yes if:
you received private income protection insurance from sources such as Aflac, MetLife, The Standard, or Guardian
you received supplemental unemployment benefits, where the employer often pays the premiums when unemployment is due to illness, injury, or a reduction in the workforce
Answer no if:
none of the above applies
16. Have you applied for or received unemployment benefits from another state or the U.S. Railroad Retirement Board in the last 12 months?
Answer yes if:
you received unemployment benefits, including extensions from any state or the Railroad Retirement Board
Answer no if:
you did not receive unemployment benefits, including extensions from any state or the Railroad Retirement Board
17. Have you been denied continued unemployment benefits because you refused to return to work or accept an offer of work at a worksite that is not in compliance with local, state, or national health and safety standards directly related to COVID-19? This includes, but is not limited to, those related to facial mask wearing, physical distancing measures, or the provision of personal protective equipment consistent with public health guidelines.
Answer yes if:
your unemployment benefits stopped because you refused to go back to a worksite that was non-compliant
your unemployment benefits stopped because you would not accept a new job opportunity at a worksite that was non-compliant
Answer no if:
you continued to receive unemployment benefits
your unemployment benefits stopped for reasons other than the above
18. Did you provide services to an educational institution or educational service agency, and you are unemployed or partially unemployed because of volatility in the work schedule directly caused by the COVID-19 public health emergency? This includes, but is not limited to, changes in schedules and partial closures.
Answer yes if:
you provided services to the above, including K-12 schools, private and public colleges; AND
you are unemployed or partially unemployed because of changes in your work schedule as a direct result of the COVID-19 public health emergency (the state considers whether you are claiming a week that is between or within academic terms and, if you are, whether you have a contract or a job is available for you to return to in the following academic year or term.)
Answer no if:
you provided services to the above but are still working or partially employed
you did not provide services to any of the above employers
19. Are you an employee and your hours have been reduced, or you were laid off as a direct result of the COVID-19 public health emergency?
Answer yes if:
your hours were reduced due to the COVID-19 public health emergency
you were laid off as a direct result of the COVID-19 public health emergency
you were laid off but could not claim regular unemployment benefits because you did not make enough to qualify, were previously disqualified, or exhausted all other benefits available to you
Answer no if:
your hours were not reduced due to the COVID-19 public health emergency
you were not laid off
as Google Documents and download in your preferred format.
User research + design kit in for improving the UI fraud reporting experience, available in English and Spanish
Download the report here.
Download the toolkit here.
USDR worked with the Customer Transformation Team at the Pennsylvania Department of Labor & Industry to conduct user research and produce wireframes that will inform a redesign of their unemployment fraud reporting tool.
The state was faced with several challenges, including:
A backlog of more than 100,000 fraud reports
Reports coming in from multiple entry points and in different formats
Reports submitted missing key information.
USDR worked in 2 phases: discovery and production of a research kit. In the discovery phase, the team focused on:
Understanding terminology, service factors, and key information flows related to fraud reporting and investigation
Identifying user barriers in the fraud reporting experience that create confusion and lead to incorrect or insufficient information needed to expedite investigations
Identifying fraud investigator pain points around processing new reports and following-up on existing reports
In the production phase, the team collaborated with the state investigation team to produce a research and interview kit. The kit will help guide the state in conducting user research, with special consideration for the sensitivity required for claimants who may have been victims of fraud and for people whose primary language is Spanish.
Reduce time to resolve newly reported fraud cases
Improve customer experience for fraud reporters and victims
Improve fraud report processing workflow
Product management
Content strategy
User research
UX Design
The team discovered 4 main pain points related to the user experience of reporting unemployment compensation fraud:
Users struggled to confidently move through the fraud reporting process, often getting stuck at the very beginning.
The website didn’t guide users to clearly provide the key information needed to efficiently investigate the fraud.
Users were frustrated by the lack of clear fraud-resolution expectations and no status updates.
Fraud reporters arrived at the reporting website feeling vulnerable and distrustful that websites can guard their personal information.
The team produced a research and interview kit for the state that includes detailed guidance in the following areas:
UX content writing principles, with a focus on plain language in English and Spanish
Fraud claimant interview kit containing:
Full Research Plan, including recruiting language, in English and Spanish
Fraud Claimant 45-Min Interview Discussion Guide, in English and Spanish
Fraud Claimant 45-Min Interview Stimulus (Figma, Google Slides), in English and Spanish
Wireframes and UX flow reviewed by investigator staff
Recommendations for how to improve workflow, measure impact, and implement a continuous improvement mechanism
The state will use these resources to conduct user research directly with people who report fraud, which will ultimately help them improve the fraud reporting experience, help them clear their current backlog of more than 100,000 fraud reports more quickly, and streamline future incoming reports.
Download the toolkit here.
Prepared by Alyssa Levitz and the Unemployment Insurance Modernization team at US Digital Response, including Waldo Jaquith and Michael Smedberg. Last updated: May 4, 2021
Most existing fraud detection in unemployment insurance (UI) systems has been aimed at identifying the claimants who are trying to claim more in benefits than they are owed, e.g., by misrepresenting their wages. Detection relied on cross-referencing the claimant-supplied information with data from their former employer(s) with government databases like departments of motor vehicles or the Social Security Administration. If any discrepancies are found, manual intervention is required.
These existing detection practices are insufficient in the face of the kind of fraud that has skyrocketed since the passage and implementation of the CARES Act in Spring 2020: identity theft.
Criminals are using stolen identities (Name, DOB, SSN, and sometimes Driver’s License ID) and using that to apply for unemployment insurance before the “rightful claimant” (i.e., the real-world person attached to the applicant’s identity) can. In this case, the identity attached to the claim is a real identity — but it is not the same as the identity of the person applying and, down the line, receiving the benefits (i.e., the applicant is a “fraudulent claimant”).
Given the changing threat model of UI fraud, quickly delivering benefits to rightful claimants with less manual intervention requires updating systems to automatically do the following:
Scan the backlog for applications that clearly are or are not fraudulent; and
Confirm that a new applicant is presenting an identity that is unique, valid, and entitled to benefits, and is themself the rightful claimant.
Confirm that the person with an existing claim who changes personal information (e.g., bank account numbers) is the same person as the rightful claimant
USDR has researched companies that provide automated identity proofing services to help in these scenarios and taken the first pass at evaluating their solutions, government compatibility, and credibility. Below is a comparison of such vendors; this list is not comprehensive, but it includes the major players and some promising start-ups. When it comes to identity proofing for workforce agencies, solutions may need to be integrated together to make a comprehensive plan. Other services may be needed for a specific use case. If you need help or advice evaluating or implementing any of these solutions, please .
By Alyssa Levitz. Last updated: Feb. 8, 2021
The vast fraud committed through the use of stolen and synthetic identities in UI programs has spotlighted the need for updated identity fraud detection mechanisms. The Dec. 2020 Continued Assistance Act mandated states make such a change for Pandemic Unemployment Assistance in particular. As States are implementing new technologies and systems, they need to consider the ways in which they are impacting racial inequities in UI benefits. States can do this by:
Finding and mitigating the ways in which their identity fraud detection flags will disproportionately affect POC
Increasing accountability for ensuring rightful claimants make it through the system
In UI systems, there is a balance between finding fraud and getting benefits out in a timely manner. Right now, a lot of fraud is slipping through and benefits are severely delayed. As we work to shift this balance, we need to ensure that we reduce -- not broaden -- racial inequities in UI.
During the Great Recession, young white men received UI benefits almost twice as often (13.6%) as young black men (7.1%). While there are many factors at play, one piece that can’t be ignored is the fact that POC are more likely to be flagged for identity fraud (though there is no evidence they are more likely to commit identity fraud).
Now is the right time to be looking at this issue, as the Continued Assistance Act of December 2020 created a new requirement for States to have digitized identity verification processes for Pandemic Unemployment Assistance. Many States had already begun doing, or actually completed this, by the time the legislation passed; however, even those States should look at their identity fraud detection processes (including the contracts they have with vendors) and surrounding user experience.
At the same time, States should also make sure that being flagged for potential identity theft does not on its own kick off other processes like trying to find benefits fraud so that the racially disparate outcomes are not compounded.
There are several common mechanisms by which people get caught up in fraud detection systems in a way that disproportionately targets POC. Below are some common problem areas and how to resolve them.
IP addresses are used to geolocate web traffic. For UI agencies, traffic coming from, e.g., out of state can be suspicious. However, there are several limitations to this approach that ultimately amount to the fact that IP addresses are not particularly trustworthy as a binary fraud detection mechanism; they should be used as part of a risk score.
IP Address databases might simply be out of date
Black and Latinx adults are 25% less likely than White adults to own a laptop or desktop computer. Without a desktop computer to file (a situation that also disproportionately happens to those who are housing insecure), you might be using devices from friends or relatives or a library (if open), or using your cell phone -- all of which means your location may change from week to week, and your apparent location may change even more.
Many UI systems were built with just English names in mind, which does not reflect the linguistic and cultural diversity of the United States. Systems need to be able to handle:
Short first and family names without error
Long first and family names without truncation
Non-Roman characters (spaces, hyphens, apostrophes, accents) without error
In addition to being able to accurately store names with the above characteristics, the identity fraud detection process needs to have some flexibility in how it interprets and cross-references names across pieces of the application. It needs to account for mistakes that the system made with someone’s name (e.g., truncated after 10 characters in part A but truncated after 15 characters in part B) as well as for cultural considerations (e.g., someone might have “Graciela” on their birth certificate but “Grace” on their driver’s license, or sometimes put a second surname in a middle name field instead of the last name field).
One common flag for UI fraud is many people filing from the same address. However, “[f]amilies of color and families with foreign-born members are more likely to live in multigenerational households,” meaning they are more likely to be flagged on suspicion of fraud. Further, the addresses of homeless shelters and other social service agencies can be used in the absence of permanent housing. This should be taken into account when determining a threshold for flagging claims.
Additionally, UI agencies should:
Ensure they are looking at all parts of the address (e.g., including apartment or floor number) when comparing addresses for multiple claims filed. Otherwise, those who live in multi-family housing units will disproportionately be targeted.
Proactively “clear” addresses known to be used by those who are housing insecure and thus could have many claims filed
Many States’ identity verification solutions require claimants to provide photo IDs corroborating their identity. This is important for a robust determination that someone is who they say they are (per NIST IAL2 standards), but it does raise some concerns.
Document verification processes can cause a number of issues from the name mismatch described above -- or a complete break in the process because someone doesn’t have the right (or any) documentation. People who are poor, formerly incarcerated, or Black are less likely than their richer, White counterparts to have a valid ID.
In any document verification step, there must be a clear way for a claimant to get in touch with a representative of the agency to find an alternative method for proving they are who they say they are. Otherwise, people without IDs are categorically going to be denied benefits in a racially skewed outcome.
All sources are linked inline. I particularly want to call out the phenomenal and in-depth research in New America's Public Interest Technology New Practice Lab's Unpacking Inequities in Unemployment Insurance report, September 2020, in particular Section 3: A Focus on Fraud Over Accessibility: The Punitive Design of UI.
Thanks also to Robin Carnahan, Sara Hudson, Waldo Jaquith, Julia Simon-Mishel, and Sam Zeitlin for the knowledge they shared with me in conversations.
A well-integrated identity proofing solution can help reduce workload and increase claim handling rates by shifting repetitive tasks from employees to automated software or outsourcing manual work from employees to vetted vendors. This can happen when the identity proofing solution is automated in a few key ways:
Identifying and correcting for errors, e.g., due to typographical errors in applications by either the claimant or employer
Identifying one-off fraud by individual applicants, e.g., intentionally duplicate claims
Identifying widespread fraud by criminal organizations, e.g., use of synthetic and/or stolen identities to claim benefits
Unemployment claims are susceptible to a few kinds of errors and fraud, some of which identity proofing can prevent or catch. Finding the right solution can be a balancing act, because you want to find fraudsters while maintaining an efficient and fair process for legitimate claims. If you tip too far one way, you may not be able to catch the fraudulent actors; if you tip too far the other way, you may incorrectly target legitimate claims, resulting in delays and hardship for claimants.
Unemployment insurance systems already have automated identity proofing solutions! They are used to cross-reference and verify the information given by an applicant and their former employer(s) with government databases like those from Departments of Motor Vehicles or the Social Security Administration.
However, more extensively automated identity proofing can help when someone’s identity is not resolved by the existing system, and can also go a step beyond by finding synthetic identity fraud. For the former, these applications are queued for manual review and additional document collection. In the worst-case scenario, this requires an applicant to mail or fax a copy of the required documents; in the best-case scenario, an applicant can upload the documents to the benefits website. An identity proofing system that automatically handles errors and uncertainties will prevent applicants from winding up in your backlog.
In the long term, issues that could be resolved by identity proofing should make up a much smaller proportion of the backlog, but for now, unemployment insurance agencies could find relief by working with an identity proofing vendor that can validate identity documentation without employee intervention.
Outside of the context of unemployment insurance, the most lucrative use case for identity proofing is in financial services, where know your customer (KYC) and anti-money laundering (AML) laws require businesses to know the real identity of their customers. Many of the vendors in this space concentrate on the financial services sector, where identity proofing is only a piece of the KYC puzzle. (Such vendors’ lack of interest in the government sector means that they are often not on schedule, but it is still plausible to sole-source their services.) However, even the identity proofing component of these services has many use cases:
Age verification: Uses identity proofing to confirm whether a customer should have access to age-limited products.
Personal safety: Dating and similar personal services use identity proofing to track customers.
Fraud detection: Credit issuers want to be sure someone is who they say they are.
Retailers: Associating online identities with persistent people enables building persistent and accurate user profiles, supporting use cases such as targeted advertising.
Note: This section has been adapted from NIST Special Publication 800-63-3 and combined with additional research. IAL2 is the level of NIST certification that is generally appropriate for unemployment insurance agencies.
“Identity proofing” is the official term for what is colloquially referred to as identity verification; identity verification is technically just a step within a larger identity proofing process. Both are umbrella terms that encompass a range of techniques to collect and resolve data to a particular person, validate that the provided data is legitimate and accurate for that person, and/or verify that the data is truly the user. At all stages, identity proofing checks the consistency of the data as it relates to a unique person, with varying levels of certainty.
Identity proofing in its most technical sense is defined by NIST, which additionally provides an Identity Assurance Level framework and certification. The IAL requirements indicate a particular level of certainty about an identity’s validity; the techniques typically associated with each stage of identity proofing are:
Identity resolution: comparing personally identifiable information (PII) provided by the user to public databases.
Identity validation: confirming that the claimant is the same person as the owner of the user account.
Identify verification: establishing a physical connection between the applicant and the PII or evidence provided.
To start with, identity proofing requires that the “self-asserted,” personally identifiable information (PII) provided by the user confirms that it belongs to a single, real person. It resolves this data by comparing it to public databases (e.g., checking the address provided by the user against a voter registration file).
A step-up in certainty would be to use knowledge-based verification (aka “KBV,” aka “Knowledge-based authentication”) to confirm the resolved identity by asking a question based on information others are unlikely to have (e.g., the system asks the user to provide the amount of their last utility bill). Depending on the specific KBV used, this technique can be a good opportunity to identify those using stolen identities.
Beyond that, more in-depth resolution techniques do not contribute toward an IAL2 designation, but they can be an important part of a solution for detecting fraud via the use of stolen or manufactured identities.
The more sophisticated forms of identity resolution are referred to as synthetic identity detection. With synthetic identity detection, machine learning (aka “artificial intelligence”) is used to combine the self-asserted data with other information you may have about a user (e.g., IP address, phone’s IMEI) and compare it with additional databases (e.g., telco records, credit header files, utility bills). It is through synthetic identity detection that criminals using stolen identities are typically found.
Some synthetic identity detection systems go even further and perform “network level” detection, e.g.:
How old is the domain name of the email address provided?
How many applications have there been from this IP address?
Have accounts on other sites been created with this combination of name + phone?
Note: a Social Security Number can be “resolved” to a person through synthetic identity detection, but it is not considered “validated” with any certainty until the Social Security Administration weighs in.
Identity validation is the first stage in NIST’s process of confirming that the claimant is the same person as the owner of the user account by evaluating “identity evidence.” With enough pieces of evidence, you can say with some certainty that a person is who they say they are. To achieve IAL2 designation, a system needs to collect between 1–3 documents and validate them with the issuing source.
Document verification is a system where a user uploads a photograph of an official document (e.g., a driver’s licence), and the validity of the document is verified through another system. If the document is verified through a system created by the vendor (e.g., that checks for accurate layout and font use, reasonable issuance dates, etc.), it is considered by NIST to be only “weak” evidence and thus does not contribute toward identity validation in the literal sense. That said, it can still be a useful feature of your system to have.
Only by checking with the department of motor vehicles that issued the driver’s license can it achieve an evidence strength high enough for use in identity validation at IAL2 standards.
Note: While a tax form could be used in document verification, using that same form to determine program eligibility is not a part of identity proofing.
Identity verification represents the highest degree of certainty that the user is who they say they are by establishing a physical connection between the applicant and the PII or evidence provided.
Common digital methods of true identity verification are:
Biometric verification is, e.g., where a person takes a selfie that is compared to their photo on an official document.
Enrollment codes or two-factor auth (2FA) is a way to verify that provided contact information is accurate by sending a code to a postal address, email address, and/or phone number (voice or SMS), and requiring that the enrollment code be provided to complete registration or login in the future.
Knowledge-based verification can be a component of identity verification by referring to data only available in authoritative and private sources.
A strong identity proofing system can be achieved either through a single product offering, or multiple products and/or vendors. When you use multiple products (either from the same vendor or multiple vendors), you can decide not to have everyone go through all identity proofing steps, and/or leverage information from multiple products at the same time to make a determination. (It is important not to chain these in such a way that would permit fraudulent identities to slip through via the weakest service.)
There are two ways that identity proofing can be integrated into a UI system for applicants to provide the required information:
Use the vendor’s API so that the applicant never leaves your application website. This option allows for greatest flexibility and can provide applicants with a more consistent experience.
From your application, direct applicants away from your website, through a process on a vendor’s website, and then back to your website again. This option is likely less work for the UI agency to implement.
Identity proofing is ideally part of the initial applicant experience. This reduces the number of applicants that need additional manual review. However, identity proofing can also be used to manage a backlog. By creating a process that encourages applicants to return to the website to provide more information, backlogged claims can be handled with less manual intervention. (Indiana did this by sending applicants a letter that directs them to an online identity quiz.)
When thinking about the backlog, there are also options that don’t require an applicant to provide any additional information. For such non-interactive identity proofing processes, the vendor will compare the applicant-provided data to authoritative data sources and return a confidence level (risk score), which indicates how certain they are that the applicant is the person who they claim to be. (Wisconsin has done synthetic identity detection on applications in their backlog to speed up determinations.)
Regardless of the way you integrate with a vendor, you should make sure that there are clear ways for someone to get direct agency help if they are unable to provide what the process requires. Doing so is an important piece of reducing the racial inequity gap of UI benefits in two key ways:
People who are poor, formerly incarcerated, or Black are less likely than their richer, White counterparts to have a valid ID. In any document verification step, there must be a clear way for a claimant to get in touch with a representative of the agency to find an alternative method for proving they are who they say they are. Otherwise, people without IDs are categorically going to be denied benefits in a racially skewed outcome.
No UX is perfect, and those with lower digital literacy may need additional support or other lower-tech mechanisms to complete their application and prove they are who they say they are. We note this here because Black and Latinx adults are 2-3 times less likely than their White counterparts to be digitally literate.
There is currently no way to tell the difference between someone who gave up because they were a fraudster and knew they couldn’t pass the step, and someone who gave up because they didn’t understand or have the required documentation to move on. An unknown number of legitimate claimants are being prevented from receiving the benefits for which they are eligible. This marks a failure of the UI system that has not received as much attention as it deserves.
Success ought to look like ensuring every legitimate claimant is able to access the benefits for which they are eligible; we should not assume that everyone who doesn't complete the application was trying to commit fraud. For more on this idea, read this OpEd or see Race and inequity in identify proofing methods.
There are many players in the commercial identity proofing & fraud detection space that are good candidates for use by UI agencies. In addition to providing key overview data for each company, we have aggregated data that will help UI agencies in making a vendor decision. We have also listed out a number of other non-functional requirements that would be relevant to most implementations.
In evaluating vendors, we came up with a list of key questions that influence the degree to which the vendor could help ease the identity proofing burden on UI systems:
What is the pricing model, and what is the cost? If it’s by verification attempt rather than only successful verification, the overall difference in per applicant cost could be 5-20% depending on the vendor’s success rate (which we don’t really know). Additionally, if a company has known set-up costs, those are noted.
What is the user experience (UX) like during identity proofing at account creation? I.e., is it an API call that is run in the background without the user noticing and/or a UX provided by the vendor that the user is sent to? Some vendors have a single product that provides an experience that all applicants would have; other vendors have multiple products that can be chained together in the “step-up” method depending on individual results.
How can it be used to process users in the backlog who have been flagged as potential fraud risks? The most impactful functionality in this area is whether they have a “batch API” that can be used to help make a determination on many individuals at once, without needing those individuals to take further action. Some vendors need a special workflow set up to send people from the backlog to their site to gather, or re-gather, information.
What methods does it use to verify identity, per descriptions in the "Process of Identity Proofing" section?
Where do they get the data against which they perform the identity proofing, including SSN? Every data source has limitations, and so in general, more data is going to result in more people with positively proved identities while continuing to catch those trying to commit fraud. (On the other hand, more data sources is likely to be reflected in a higher price for that vendor.)
If you believe that a vendor is a good match for your needs based on the key considerations above, the following information about the company could help you finalize your decision. Please do not be discouraged by “unknown” answers for some of these questions — that we have been unable to get answers to these questions does not mean that they are unanswerable.
What notable (name-brand) customers do they have?
What special certifications/authorizations does it have?
Have other government entities used it?
Does this vendor have existing contracts through an available Federal Supply Schedule (FSS) through GSA or some other Governmentwide Acquisition Contract (GWAC)?
Is this vendor under a recognized socioeconomic program or status such as the 8(a) program or Service-Disabled Veteran-Owned Small Business (SDVSOB)?
While most state unemployment insurance agencies are trying to solve the same set of problems, the technologies and processes that they are working with vary greatly. Each organization will have to determine their own requirements in the following areas:
Network API and style (e.g. REST/SOAP/GraphQL)
Supported development languages
Client libraries
Security concerns
Average and 99th percentile response times
Scalability
Error rate
Support (API docs, consulting services, third-party support, etc.)
Licensing, embedding, reuse
Data storage and access policies (Do they store PII on their side? Do their employees have access to that data? If so, how are those employees vetted and/or held accountable?)
Severability and replaceability (How is their contract structured? If they store data, is that data accessible in a bulk, machine-readable format?)
Hosting model (SaaS or on-premise?)
For each of the evaluated vendors, we have done our best to provide accurate information through a combination of research and conversations with company representatives. This section of the document summarizes some of the more important vendor differences to inform your decision making. For all the details, see the appendices:
There is an identity proofing vendor that falls outside of the scope of this document, but that is likely to be of interest to readers: Login.gov, provided by the federal government. The single-sign-on service was launched by the General Service Administration in 2017, providing two-factor authentication, fraud detection, and Identity Assurance Level 2 (IAL2) under NIST-800-63A. It was initially available only to federal agencies, with a FedRAMP Moderate ATO, with customers including the Department of Defense, the Department of Homeland Security, the Department of Energy, and the Department of Transportation. At the end of 2020 they were granted permission by the White House Office of Management and Budget to accept state agencies as customers.
Login.gov is not a drop-in identity proofing vendor. They perform identity proofing, but only as a component of a user registration process within Login.gov. For employment agencies to use Login.gov for identity proofing, they need to replace their entire authentication flow with Login.gov, integrating it via OAuth 2.0 or SAML.
Name, website, and last date updated in this document
Headquarters
Founded
User Base
Best for
New York, NY
2015
Financial services, banking
KYC/AML compliance, fraud prevention
Palo Alto, CA
2014
Financial services and marketplaces
KYC compliance; address and age verification
Seattle, WA
2012
Online lending, retail banking, ecommerce and marketplaces
Identity records for dynamic PII
Dublin, Ireland
1996
Government partners, financial services, online lending
KYC compliance, fraud prevention, identity records
McLean, VA
2010
Government partners, retail, online healthcare
Identity records
Idemia
https://www.idemia.com
5/4/2021
France
2007
Government partners
Identity records
Tallahassee, FL
2003
Financial services, banking, retail
Identity and age verification
San Francisco, CA
2017
Retail banking, credit card issuers, all types of lenders, and fintech
Synthetic fraud detection & analytics
New York, NY
2012
Retail banking, credit card issuers, and remittance providers
Fraud scoring and analysis
Alloy is the most configurable of the vendors; they have partnerships with many other vendors that provide a wide variety of identity verification methods that can be used in combination with each other. Their partners include most of the vendors evaluated in this report: Cognito, Ekata, IDology, SentiLink, and Socure. We do not have information on their pricing.
Cognito’s identity proofing focuses on basic PII: Name, phone, address, and SSN. Their unique offering is through using synthetic identity detection to confirm the validity of a Name / Phone number combo, and then using 2FA to confirm that the person is still in possession of that phone number. (Additional KBV is an add-on for further detection of stolen identities.) With this reliance on 2FA, their product isn’t as suitable for managing the applicant backlog without needing the applicants to take some action.
Ekata specializes in confirming “dynamic PII” -- Name, phone, address, and email. By also looking at passively-collected information (e.g., IP address and phone metadata), they are able to detect stolen as well as synthetic identities. They do not have a batch way to process the applicant backlog without needing the applicants to re-enter this basic PII. They do have a dashboard where you can see the results of an individual’s ID proofing process.
Experian is one of the vendors that could provide all the identity proofing pieces; they have both a step-up offering and a full NIST IAL2 offering. It is one of two vendors that appear to have gotten contracts with state UI agencies since the passage of the CARES Act (5-6 states). Their synthetic identity detection product can be used on the applicant backlog without needing the applicants to take any action (document verification would of course need the applicants to provide that documentation). They were the one company that mentioned the use of “marketing data” as one of many data sources used in their synthetic identity detection.
Idemia is another of the vendors that could provide all the identity proofing pieces, partially through the way it leverages Experian Precise ID. They appear well set-up to be used in a “step-up” identification process for either applicant creation or backlog management (i.e., their batch API can do synthetic identity detection without additional action from the applicant). If a state's Department of Motor Vehicles uses Idemia, that state's UI system can be configured to use that database as another trusted source.
IDology is another of the vendors that could provide all the identity proofing pieces, though we don’t have their pricing information. They appear well set-up to be used in a “step-up” identification process for either applicant creation or backlog management (i.e., their batch API can do synthetic identity detection without additional action from the applicant). Part of their unique offering is access to the Consortium Fraud Network that allows them to securely check the use of PII combinations in additional contexts.
SentiLink focuses on synthetic identity detection, comparing the self-asserted PII to numerous data sources. The breadth of their data sources means that with sufficient PII collected, they should be able to detect stolen identities; however, they did not mention the use of passively-collected information, which can be very helpful in this regard. They can be used at either application creation or to evaluate applicants in the backlog without the applicant needing to take action; they also have a dashboard where you can see the results of an individual’s ID proofing process.
Socure is another of the vendors that could provide all the identity proofing pieces. They explicitly recommend creating a “step-up” process and shared that the synthetic identity detection step can verify 90% of people, leaving only 10% to need the more expensive doc + bio verification step. They can be used for either applicant creation or backlog management (i.e., their batch API can do synthetic identity detection without additional action from the applicant)
Given the variety of products from these vendors and the way that the information was provided, it’s hard to do a direct comparison -- but we can try by making a couple of assumptions and establishing some constants:
For the vendor with a set-up fee (Experian), the cost is amortized over 2 years and 10,000 claims per month.
For the vendors that provided approximate costs:
Sentilink said $0.25 / verification: we will create a moderate cost range of $0.15 - $0.45 / verification.
Socure said mid-to-high single digit cents / query: we will create this as a range of $0.04 - $0.09 / query.
Not all queries will result in a verified identity; we will stipulate that a best-case scenario for all vendors other than ID.me is 95%, and that the worst-case scenario is 80%.
*Because ID.me has a virtual in-person proofing step that none of the other vendors have, we will stipulate that their best-case scenario for achieving a verified identity is 99%, and that their worst case is 85%.
Below are the results of calculating average cost per query (i.e., cost per new UI claim) for the 6 vendors whose pricing information we have
ID.me is another of the vendors that could provide all the identity proofing pieces; their primary offering is a full NIST IAL2 identity proofing solution. (They also offer pieces as individual products, but we do not have as much information on that.) It appears to be the primary vendor that has gotten contracts with state UI agencies since the passage of the CARES Act (AZ, CA, CO, FL, GA, ID, IN, LA, ME, MA, MS, MO, MT, NV, NJ, NY, NC, ND, OR, PA, SC, TX, WA). Their document + biometric verification solution is the most sophisticated; if someone cannot be verified through a comparison of a selfie to the uploaded documents, they are routed to a “remote in-person” identity proofing video chat where those documents are presented in real time to an ID.me call center. There have been concerns around the wait times of this service, To be used in applicant backlog management, it requires that everyone be sent to their site to re-enter their PII and provide documentation because it is an IAL2 certified solution.
Low average cost per query | High average cost per query |
Cognito | $0.56 | $0.94 |
Ekata | $0.10 | $0.25 |
Experian | $0.11 | $0.28 |
ID.me | $3.40 | $3.96 |
Idemia | $2.00 | $5.23 |
SentiLink | $0.12 | $0.43 |
Socure | $0.09 | $0.29 |
Alloy and Experian are different from the other vendors in that they explicitly and transparently leverage other companies’ technology as identity proofing platforms. Both take information that others have already interpreted and use that as an input to their own risk interpretation. Additionally, Idemia itself uses Experian.
Alloy has a platform that can combine multiple risk assessments about the same piece of data. For example, Alloy can interpret in parallel the fraud risk data sent by both SentiLink and Ekata about a particular Name + Address combination. Alloy has agreements with more than 65 partners in total that can be mixed and matched when setting up an identity proofing system. This immense flexibility (and overlap with the vendors evaluated in this paper) and the fact that no pricing information was made available make it hard to come to any conclusion about Alloy.
Experian’s approach is slightly different. They build some of the infrastructure themselves, and they rely on other companies for specific pieces of the puzzle: Acuant provides their document verification; EmailAge by LexisNexis gives them a risk assessment specifically about the longevity of an email address and the domain to which it belongs. Experian bundles the vendors and features into 2 or 3 offerings, as distinct from Alloy’s a la carte approach.
UI agencies themselves could also build their own platform by using different vendors at different steps of the process. (If you wanted multiple products to provide synthetic identity detection on PII, it would likely be more effective and less risky to achieve that through Alloy--with the caveat that their pricing is unknown.) For example:
Use your existing method to determine validity of SSN / Name / DOB combination
If that combination is valid, collect and use additional PII (address, mother’s maiden name, email, phone, etc.) and evaluate it with one vendor’s synthetic identity detection product.
If step 2 indicates a particular fraud risk, have that individual go through a document + biometric verification step.
UI agencies already have in place methods to determine whether a provided SSN / Name / DOB combination is real (i.e., not synthetic). That functionality can remain in place alongside any new identity proofing mechanisms, as long as there is clear communication between pieces of the system.
There are 2 ways that vendors determine the validity of a provided Social Security Number (Ekata doesn’t handle SSNs at all). The second method is more “official,” but both Cognito and Socure believe their method to be effective.
Cognito and Socure have systems that search for prior use of a SSN / Name / DOB combination, e.g., through DMV records or credit files. A drawback of this approach (depending on the details of the vendor’s implementation) is that it is possible for synthetically created identities to have credit files. Additionally, the vendors that rely heavily on credit bureau sources will systematically be less likely to prove the identities of those with less access to credit. Via both Cognito and Socure, Alloy has this functionality.
Experian, ID.me, and SentiLink have systems that check against the Social Security Administration Death Master File, which will not have any synthetic identities in it. All three vendors use additional methods to detect synthetic identities. However, it is updated at most weekly and is not a comprehensive record of all deaths in the country; a notable exception is that it excludes state death records. Via SentiLink, Alloy has this functionality.
Note: Two vendors, SentiLink and Experian, have access to the SSA's new eCBSV product, which for their financial customers only allows them to effectively query the SSA directly and thus know with near-certainty that the SSN / Name / DOB match, and belong to a live human. Unfortunately, the SSA does not currently allow for other uses of this API, despite how useful and effective it would be for the UI identity theft detection scenario.
Digitizing the document verification step for those whose identities are in question is a key part of detecting fraud and of reducing load in UI agency staff. As discussed in Recommended Process for UI Identity Proofing, the system can require everyone to go through a document verification process, or only a subset of individuals (a “step-up” process).
The following vendors provide document verification:
Alloy
Experian
ID.me
Idemia (via Experian Precise ID)
IDology
Socure
All of these vendors’ documentation verification products combine it with a biometric verification step. The biometric verification is in the form of a selfie that gets checked for “liveness” (i.e., to check whether someone else’s existing photo was uploaded) as well as compared against the provided photo ID.
ID.me’s document + biometric verification is built into their primary product offering. They take doc + bio verification a step further than any of the other companies with a "remote in-person” identity proofing interview: a video call in which an applicant must present their documents live. This happens only for the set of people who cannot be verified at other steps in their process. (You can also get their document verification and/or doc + bio verification products individually, but we don’t have information about the pricing.)
Note:
ID.me is planning on adding 500 in-person identity proofing "outposts" beginning in June '21
Idemia's IdentGO product could also be used by states looking to outsource physical/in-person identity proofing
IDology, and Socure have their own stand-alone document verification products that could be used as the “step up” from a different vendor’s synthetic identity detection. Alloy and Experian could likely provide just a document verification service, but it may not be the most efficient way to do so: Both use Acuant to provide document verification.
The following vendors do not have document verification:
Cognito
Ekata
SentiLink
USDR can help you further evaluate vendors to find out which will best suit your state’s specific needs. Reach out to ui-team@usdigitalresponse.org if you are interested.
Follow 18F’s De-Risking Government Technology: State Field Guide when planning your identity proofing automation project.
See “Appendix B: Acceleration Plan for Identity Verification” in California’s Employment Development Department Strike Team Detailed Assessment and Recommendations to serve as a reference for how you can create a plan for your state.
If you have questions or need help solving unemployment insurance issues in your state, please contact the Unemployment Insurance Team by filling out this form.
For more information on each of the considerations, please see Key Considerations.
For some of the vendors, the answer is “unknown,” and we continue working to try to find that information. For ease of reading on a variety of screen sizes, the evaluation is split into two charts:
Pricing
UX at account creation
Backlog management
Alloy
Unknown
Step-up process:
- API for PII collection
- UX for doc verification
- Has batch API to check PII
- UX for doc verification
Cognito
$0.70 - $0.99 / verification
API for PII collection
Has batch API available but won’t be able to verify phone number
Ekata
$0.10 - $0.25 / query
API for PII collection
No batch API to check PII
Experian
$0.10 - $0.25 / query, plus $2,500 or $6,500 set-up fee based on implementation detail
Step-up process:
- API for PII collection
- UX for doc verification
- UX for UI agents to input info provided via phone
- Has batch API to check PII
- UX for doc verification
ID.me
$4.00 / verification; $2.00 for subsequent years’ renewal
- Complete UX for PII collection and doc verification
- UX for PII collection and doc verification
Idemia
$2.50 to $5.50 / verification
Step-up process:
- API for PII collection
- UX for doc verification
- Has batch API to check PII
- UX for doc verification
IDology
Unknown
Step-up process:
- API for PII collection
- UX for doc verification
- API for PII collection
- UX for doc verification
SentiLink
~$0.25 / verification
API for PII collection
Has batch API to check PII
Socure
Mid-to-high single digit cents / PII query; add’l $1/query for doc verification
Step-up process:
- API for PII collection
- UX for doc verification
- API for PII collection
- UX for doc verification
Identity proofing methods (and product names, if applicable)
Data sources
Alloy
Highly configurable: KBV, 2FA, synthetic identity detection (including looking at device “fingerprinting” and IP address), document verification, biometric verification
Partnerships with 65+ data vendors including ones we have or will evaluate: Acuant, Iovation from TransUnion, Ekata, Socure, Cognito, Sentilink, and IDology
Cognito
- Cognito Identity Verification Service: 2FA, synthetic identity detection
- Blocksore product: 2FA, synthetic identity detection, KBV
Credit bureaus and public data sources
Ekata
Synthetic Identity Detection
Public data sources -- does NOT check SSN
Experian
Experian PreciseID: Synthetic Identity detection, KBV, 2FA
Experian Identity Proofing: Document verification, biometric verification
Credit file info; SSA Death Master File; marketing data; Motor vehicle info from auto dealerships and DMVs; public data sources
ID.me
Synthetic identity detection, document verification, biometric verification
Credit bureaus, SSA Death Master File, telco records, public data sources
Idemia
Idemia Proof: Synthetic identity detection, document verification, biometric verification
Credit file info; DMV database; "Systems of record"
IDology
ExpectID: Synthetic identity detection
ExpectID IQ: KBV, synthetic identity detection
Document Scan Solution: both document verification and biometric verification
Motor vehicle info, voting record, creditor information, utility bills
Have “Consortium Fraud Network” for additional commercial sources
SentiLink
Synthetic identity detection
Credit bureaus, utility data, SSA Death Master File, selective service information
Socure
KYC identity verification module & Sigma Fraud products: synthetic identity detection
DocV product: both document verification and biometric verification
Credit bureaus, utility data, telco, public data sources
An e-adjudication case study
Note on U.S. Digital Response Independence
U.S. Digital Response’s (USDR) unemployment insurance (UI) team has partnered with more than ten workforce development agencies to troubleshoot challenges, expand capacity, and improve the user experience. With our partners, we quickly evaluate, plan, and/or implement more effective and impactful solutions. Through one of these partnerships, we learned about their newly digitized fact-finding system. USDR was not involved in any way in the project described in this case study. We wrote this case study because it is a great example for other agencies of a UI modernization project that uses incremental, continuous delivery, and modular development while keeping the mission in-house.
Impact Summary
This State UI System was slowly resolving claimant issues because they were fully reliant on fact-finding via phone calls scheduled weeks into the future. The State implemented a new system, switching to an online fact-finding interview, resulting in a response from claimants within one day for 50% of issues, instead of weeks for each individual concern. In moving to an adaptable new system, examiners nearly doubled the number of cases they were able to process in a day.
Problem Statement
Solution
We incorporated USDR’s fundamental principles in the following ways:
Modular development: the State DOL built the system “on top” of the existing infrastructure, to replace the existing manual phone-based adjudication system. They do have API-based integration from the mainframe to initiate the new adjudication system, but they still have examiners input determinations directly into the old system rather than requiring that the new system connect directly back.
Incremental, continuous delivery: The State DOL did not transition all issues to the new system at once; they started with issues connected to initial claims, then progressed to web certifications/weekly claims, reopened claims, and monetary issues. This iterative process provided time for staff to get used to the new system and to improve the process for some issues while the system was being built out for others. They didn’t have to wait for everything to be done to start seeing benefit to both staff and claimants. In addition to not waiting for the system to cover all issues before going live, the State DOL kept several features out of scope and can consider adding them in the future, e.g.: employer fact-finding, carefully automated determinations, API connection back to mainframe from the CMS.
In-house ownership: Their vendor showed the State DOL how to create and edit questionnaires, notifications, and reports, so, once the system was initially set up, they could easily make their own changes or additions. The State DOL is able to use the reports to understand what claimants might be having difficulty with in the user experience and make copy and question changes accordingly.
Architecture
User Experience
Once the CMS knows which claimants have what issues that need resolving, it sends email notifications to claimants with unique links. The content at this link is specific to the individual, and can gather information about both monetary and non-monetary issues.
The CMS can be set up to send reminder notifications for people who haven’t responded in the given timeframe; the State chose to leverage both email and SMS capabilities for reminders. The reminders can be tailored to different populations, e.g., those who haven’t started the questionnaire, who started and abandoned the questionnaire, or who submitted the questionnaire with incomplete responses.
When the claimant fills out the questionnaire, their case is put in a queue for examiners to look at. The CMS enables load balancing cases for the examiners (and can account for differing skill levels/proficiencies), and provides reporting on productivity and outcomes for individual examiners. An examiner can see data about all the claimant’s issues in one place, and then when a determination is made, they enter that information both into the CMS and manually back into the mainframe.
The State planned for a 6-week period where phone interviews were still happening, but new ones weren’t being scheduled. During the transition period, examiners needed to handle both the scheduled phone calls and the new digital workload. It was not as bad as everyone thought it would be given the efficiency and ease of the new system. They went live based on when they knew they would have lowest volume, and they were prepared to turn the new system off and keep scheduling phone appointments while troubleshooting before doing a small live test again.
For more information on each item, please see Supplemental Considerations.
Note: as none of the evaluated vendors are under a recognized socioeconomic program or status, that information isn’t duplicated in the chart below. For some of the vendors, the answer is “unknown,” and we continue working to try to find that information.
Previously, the State UI System was far from meeting the US DOL target of 87% of issues resolved within 3 weeks, and were put on a corrective action plan. At that time, to resolve issues the State DOL was fully reliant on scheduling phone calls with claimants, creating a backlog so large that calls needed to be scheduled weeks into the future. To become compliant with the U.S. DOL’s standards, the State would need to drastically change their fact-finding process.
The State DOL stood up a cloud-based case management system (CMS) to get more information from claimants when issues are detected. This new CMS is nearly independent of the State’s mainframe: a query is run nightly in the mainframe to get a list of claims with particular issues, and that data is sent via API to the CMS to send questionnaires to claimants. On the other end, examiners view responses in the CMS and do double manual data entry into both the CMS and mainframe.
Notable commercial customers
Notable government customers
Relevant certifications
Schedule information
Alloy
Austin Capital Bank, Langley Federal Credit Union, Radius Bank
Unknown
Unknown
Unknown
Cognito
Brex, Nextdoor, Coinbase, BBVA
None
SOC2 Type 2 Compliance
No
Ekata
Lyft, Alaska Airlines
None
SOC2 Type 2 Compliance
No
Experian
NASWA Integrity Data Hub
Healthcare.gov, IRS, Michigan's MyLogin, 5-6 state UI agencies
NIST 800-63-3 IAL2 (for their full ID proofing product)
GSA (GS-35F-188AA) & NASPO
ID.me
LinkedIn, Lenovo
Vets.gov, 22 state UI agencies
NIST 800-63-3 IAL2/AAL2; in process of FedRamp authorization
Have a growing government business line, but do not promote availability on any found schedule
Idemia
n/a
TSA, USPS, US Dept. of State, 37 state DMVs, 1 state UI agency (Okla.)
-In process for NIST IAL2
-IEC CD 18013-5 Compliance
GSA
IDology
Unknown
Unknown
Unknown
Unknown
SentiLink
Several of the largest banks, credit card issuers, credit unions, and auto lenders
With at least one state’s PPP
SOC2 Type 2 Compliance, PCI Compl., EI3PA Compl.
Unknown
Socure
Seven of the nine largest U.S. banks, six of the top 10 U.S. card issuers, Chime, SoFi
Unknown
-In processes for FedRamp authorization & NIST IAL2
-SOC2 Type 2 Compliance - ISO 27001/ 27017/27018
Available on AWS Marketplace, so it may be simple to procure via an existing AWS contract
By Alyssa Levitz
Phone lines, or other high-touch customer support channels like in-person appointments when safe, are an important part of the UI system; there are some kinds of issues that can only be solved via a synchronous, human-to-human conversation. They are even more important for those with lower digital fluency or less internet access.
Black and Latinx adults are 2-3 times less likely than their White counterparts to be digitally literate. Having a better user experience, including the use of plain language (both in English and when translated), not only increases the likelihood that everyone can effectively navigate the UI system on their own, but it also reduces the need for high-touch customer support channels for those with higher digital literacy. Chatbots and other digital self-services to answer questions or resolve issues would be similarly helpful.
The improved UX needs to include clear designs for mobile devices. Black and Latinx adults are 25% less likely than White adults to own a laptop or desktop computer, an inequity which is compounded by the closing of libraries and other public spaces where such devices are typically available.
For the 14 things you can do that will most improve your UI system's user experience, see Section 5, Stage 2, Recommendations 2.3 of: "Centering Workers—How to Modernize Unemployment Insurance Technology," by The Century Foundation, NELP, and Philadelphia Legal Aid; October 5, 2020.
For a deeper understanding of how digitizing processes is important yet can deepen racial inequities -- and considerations for preventing that, see: Section 6: The Digital Divide & UI Modernization: States' Moves to Online Applications Worsen Accessibility from New America's New Practice Guide's Report on Unpacking Inequities in Unemployment Insurance.
By Alyssa Levitz. Updated: Jan. 25, 2021 to note CT and KY have added this functionality
They have implemented an in-house call management system, focused on call scheduling. Here’s how they describe the current system:
Customers call the main claims center line. A triage agent handles basic triaging and addresses the needs if possible. If the need is more complex, then a callback is offered. The callback is scheduled by the agent via webpage using a tool from a company that specializes in event scheduling. Available time slots are verbally presented to the customer
In addition, they have the option for a self-service page where claimants can schedule their own appointment using the same method offered to the triage agent. This self-scheduling is currently turned off, as there was too much demand compared to allocated staff, and there were too many duplicate appointments being made.
Each day, agents look on a separate webpage that has all the callbacks listed. Agents have basic information collected by the triage agent to pre-scan before making the call.
In a survey of other states’ Departments of Labor / Unemployment Insurance Departments, we found two with easily accessible online appointment/callback scheduling.
Colorado: Leverages YouCanBook.me (https://youcanbook.me/): see it embedded at the bottom of their contact page even when appointments don’t exist. They do not mention when new appointment slots are made available, but none were available every time we’ve checked.
Michigan: They use Acuity (https://acuityscheduling.com/): They also have a separate page for the scheduler; unlike Alabama and Georgia, it still lets you start the process of finding an appointment slot even if none exist. They say, “New appointments become available throughout the day exactly 7 calendar days in advance (for example, on a Tuesday morning appointments for the following Tuesday morning become available).” None have been available any time we’ve checked.
Two other states have systems in place that seem to target the same solution area: the phone lines are long and unpredictable, so having known times makes the experience better for some.
Alabama: Alabama’s tool takes basic constituent information and assigns them a time and location (theirs is not for callbacks but actual in-person appointments). You can see it on this page, but it’s unlikely to show you much unless it’s a Friday afternoon at 5PM CT when they’ve opened up new time slots.
Georgia: It is not clear what Georgia’s functionality is or how they’ve implemented it, as no appointment slots have been available on the separate page they built. It’s clear that going through that flow will result in some kind of appointment made, but not apparent how that appointment time is chosen/set. New appointments are released Monday mornings at 8 AM ET.
Since this survey was initially published in early Dec. 2020, Connecticut and Kentucky have added online appointment scheduling, but those have not been looked into.
Potentially interesting to you is how state Departments of Motor Vehicles or Departments of Licensing have implemented their online scheduling systems, as their use case is similar to UI agencies (e.g., the need to collect enough information to know who a constituent is, there’s a known number of appointments available per time slot, and it doesn’t matter within a group whom a constituent speaks with).
Vendors in this space with known government contracts are:
Acuity (https://acuityscheduling.com/): Michigan’s Secretary of State
Nemo-q (https://nemo-q.com/): Georgia, North Dakota, and Utah state licensing departments; Washington, D.C.’s DMV; and Florida’s Broward, Miami-Dade, and Volusia counties’ departments of transportation
Qmatic (https://www.qmatic.com/): New York’s Monroe County
Vcita (https://www.vcita.com/): Hartsfield-Jackson: Atlanta International Airport; Florida’s Osceola County’s Circuit Court Clerk; Hawai’i’s Kaua’i’s and North Carolina’s DMVs; Kansas state department of Revenue;
Additional vendors with DMV-targeted offerings:
Setmore (https://www.setmore.com/dmv)
Q-less (https://www.qless.com/)
States with unique-seeming solutions for their DMVs:
Unknown: California, Indiana, Oregon,
Tool referred to as “Fortress”: Washington
There are three ways that the Candidate vendors price their tools: (1) per month, and (2) per month but different depending on the number of users, and (3) per user per month. Depending on the particulars of how the appointments made via this tool get surfaced to frontline agents, the vendors that charge per user per month could become astronomically expensive.
(1) Per month:
YouCanBook.Me: Their rate is actually per calendar per month, but given that you’d create some (likely small) number of calendars up-front, it’d be a known cost per month unrelated to the number of users you have. (Note: they might have discounts for government.)
Nemo-q: Their rate depends on the number of appointments you want to schedule a month.
(2) Per month, with a limit on the number of users at each pricing tier (i.e., the tier you choose is influenced by both functionality and the number of users needed -- but the cost doesn’t scale linearly with number of users):
Acuity
Appointy
FlexBooker
Vcita
(3) Per user per month (i.e., the cost scales linearly with number of users): This pricing set-up appears to be because the tool assumes that every agent would have a calendar within the product that gets synced to the “meta” calendar.
Calendly (Note: they might have discounts for government.)
Microsoft Booking (not actually known for certain that they charge per user per month, but it is the most likely case)
Microsoft Dynamics
Vcita (Note: they have discounts for government)
The evaluated vendors fall into three main categories, only one of which feels like a good match for the state UI agency's scenario: “General Appointment Scheduling,” which is also referred to as “Candidate” to reflect that match.
Healthcare appointment scheduling: These ones tend to have rich feature sets and good UX for the actual scheduling mechanics -- but are not as applicable in other ways that the agency would need. E.g., they may have only medical-specific intake forms (like uploading insurance information), or they don’t have integrations with CRMs (instead integrating with Electronic Health Record management systems). Several of them were also more focused on queue management than the actual scheduling of appointments.
Small service-sector business appointments: These vendors are targeting small businesses with a handful of employees for 1-1 or group appointments usually with a SPECIFIC individual. They also tend to offer a number of things that the agency wouldn’t use (like social media integration, payments integration, and marketing features), contributing to a higher price point than some of the other offerings.
General appointment scheduling: These vendors are targeting organizations or even individuals across sectors. Their systems are generic enough to be useful in a lot of scenarios, including the UI agency's, and so in this paper are referred to as “Candidates.” Some of these vendors have solutions targeting state Departments of Motor Vehicles or the equivalent, which is a useful comparison scenario for these purposes.
Microsoft Bookings: While the actual UX may not be that customizable, it has so many options, particularly around defining “service offerings” (what we would call “appointment types”) that would help it meet the needs of the organization.
Runners up: Both Vcita and Nemo-Q, with their targeted DMV offerings, appear to have all the configurability needed.
YouCanBook.me: This vendor does quite well with the various requirements against which we evaluated all the vendors, and its pricing model (per calendar per month) makes it the cheapest option. With the caveat that we're sure prices are negotiable, we're going to use an example to show why YouCanBook.me appears the best on price.
Say you wanted 3 calendars; it’d be $30/month (we think your use case could actually be served by appointment types within the same calendar, but we are using multiple calendars to make this comparison a little more starkly. To beat that monthly price, each vendor would only let you have the following number of users:
Acuity: 6 users, their “Growing” tier at $23/month
Appointy: 1 user, their “Growth” tier at $19.99/mo
Calendly: 3 users, their “Premium” tier at $8/user/mo; or 2 users, their “Pro” tier at $12/user/mo
FlexBooker: None
Microsoft Bookings: Unknown
Microsoft Dynamics: None
Nemo-Q: None
Vcita: 1 user, and it’s the “Essentials” tier that doesn’t have the functionality you’d need
Calendly: This was a bit of a hard one to determine without a sales call and demo, but at least from how their scheduling system was described, it seems like their “pooled availability” and/or “round robin scheduling” systems would be really great fits for what the agency is looking for -- and none of the other vendors at least so explicitly outlined a part of their service that would meet your needs.
Runners up: Any of the vendors with per-user calendars. It really depends on how you currently manage staff calendars and how you’d want that to integrate with the appointment scheduling solution.
By Alyssa Levitz. Updated: Dec. 22, 2020
One common channel for customer support is some form of synchronous online "chat." The National Association of State CIOs published a document about the broader use of chat by states during the pandemic; this document dives specifically into chats set up to support UI agencies: what types of chat do they use, and what are some of the best practices we see?
The target audience of this report is primarily state UI agencies that are looking to add or revamp their online chat offerings. Much of what is discussed here, though, is applicable to many other government use cases.
Prompted chat in which users are guided through a series of pre-defined questions and answers is likely to be most effective, but whatever chat type you use: make sure to set users' expectations up front.
Accessibility is a key consideration for chat experiences and should be incorporated from the beginning of the process.
Use data to inform the initial set up of your chat experience, and make sure to have data collection in place to iterate and improve over time.
Your system should handle 15-20 questions or scenarios very well and have clear paths for people to find more information on other topics.
We looked at each state's UI agency government website, and in some cases non-governmental but official websites that people would be directed to in order to learn about or apply for or manage unemployment claims. In many cases, we took screenshots for easy reference for myself; this also has the effect of documenting some changes over time.
Once a chat functionality was found, we used the Chrome browser's Developer Tools to find evidence of which vendor was used to provide it. With this method, we were unable to determine the vendors used in Kansas, Massachusetts, and Maryland. Vendors may have more chat-related offerings than are mentioned here; this document is not intended to be a comprehensive vendor analysis.
As it relates to UI agencies, online chat implementations fall into one of four categories, which I'm calling: Live, free-text, prompted, and support. Some states have multiple implementations on different parts of the site or at different times of day; others have multiple and let the customer choose which one suits their needs. For each type of chat, we will discuss its pros/cons as well as vendors seen to be offering that functionality to UI agencies as of December 14, 2020.
Links to how you can access each state's chat system, along with any notes like what vendor they used, can be found on the page where we're tracking solutions across many states:
Live chat is akin to instant messaging; there is a human on both sides of the interaction. It can be set up in a multitude of different ways; the most important distinction among them is what the chat agent is able to do: Can they answer only generic questions and point people toward existing documentation? Can they help with password resets? Can they answer specific questions about someone's claim?
Pros:
Humans can interpret free-text responses better than any machine learning natural language processing model, so people are less likely to be left without a way to at least begin to get their question answered
More easily than the other methods, live chat can be set up to answer questions about individual claimants' concerns
Cons:
Still constrained by number of employees or contractors, the learning curve associated with training them, and how much system access they are given to answer people’s questions → i.e., people get put in queues, there isn’t 24/7 availability
Vendors:
Genesys
MedChat
SalesForce
Availability: For your chat system to have the most impact, you should make sure that the chat is available on any/all pages of your site (or sites!) that get meaningful traffic by people who might be looking for help with unemployment insurance. This would include any separate sites you might have for:
UI Agency administration
Standard/PEUC UI management
PUA UI management
Re-employment services and support
Some design basics:
Don't call it "live" if it's a bot or virtual agent, either in the chat itself or in content that references it. This sets the wrong expectation for people and might lead them to be more likely to provide private information even if it is not appropriate for them to do so.
The first message from the system should be short and to the point. Make sure that it is fully visible within the interface without the customer having to scroll. This means that the chat's "header" should be relatively narrow, and that there aren't large images or large blocks of text.
If content is split into multiple messages, then leave some time in between message sends for the person to at least register that there are multiple messages, if not fully read one message before the next one gets sent
For prompted chats, make it easy for users to get back to the "main menu" so that people can get multiple questions answered or change their mind about an answer response.
When implementing a chat or chat-like experience for your UI agency, you need to prepare a list of likely questions and acceptable answers. This Q&A list is used to build the decision tree and responses for Prompted Chat, to prepare agents for Live chat, to prepare Free-text chat to understand a variety of inputs, and to provide answers for all those chat types as well as "support chat.
You should prepare your system to handle 15-20 questions or scenarios very well, and have clear paths for people to find more information on other topics. There are many kinds of data you can leverage to inform this initial Q&A set, for example:
From your IVR system: what branches are most commonly used?
From your call center: what are common topics of calls? What questions are they answering over and over again?
From your claims agents: what mistakes are they seeing people make, or misunderstandings people have?
All of this data is useful in other ways, too, as you look to make more fundamental improvements to your UI system. E.g., if people frequently ask about their claim status, you could consider prioritizing work to make that possible online (or if it's already possible, to move it to a more expected location or change the language around it to make it clearer).
Regardless of the type of chat, you should make sure to have metrics in place to track the effectiveness of the tool and make changes moving forward. As economic conditions, policies, and your benefits website change, the chat experience needs to be kept up-to-date.
Effectiveness measures:
CSAT for tracking people's satisfaction with the response they got
After the chat thinks it has done its job, you can have it ask, "Did that answer your question?" For making changes, you should keep an eye in particular on the sequence of back-and-forths that lead up to someone's question not being answered - though don't forget to celebrate all the times that it did answer someone's question!
"Engagement" metrics: e.g., button clicks, messages sent, clicks on outbound links
Time spent: for chats other than live chat, how much time are people spending interacting with the tool?
Language analysis: for chats that let users input any text they want - what are they asking for that the system doesn't understand?
"Free-text" chat is when a user types their own message. The system then uses Natural Language Processing to try to interpret the message in order to provide the most relevant pre-written response. This is what people most commonly think of when they hear about "chatbots."
Pros:
People can submit any question
Cons:
People may get more frustrated at the ability to ask any question but not be able to get all questions answered.
The system may not be able to interpret what someone is saying
Vendors:
Astute Chatbot
Microsoft
Zendesk
"Prompted" chat is when people are presented with a series of options for what they can get help with. The options are frequently structured as a decision tree, with sub-questions for the primary scenarios. This chat seems to always come with f-text chat, so you can get all the pros (and cons) of that, though the intended primary interaction is for the user to click buttons within the chat window rather than typing anything into the chat box.
Pros:
You can configure the flow to answer the most common questions
People have clear expectations of what they can get help with
Content is created specifically with the chat format in mind
Cons:
Without careful set up, it is very easy to create "dead ends" for people seeking help.
Vendors:
AWS/Accenture
IBM
Microsoft
Salesforce
Twilio
"Support" chat is when a chat-like interface is set up to help someone navigate available help articles that are likely seen elsewhere on the site, e.g., linked to from an FAQ page. When someone types into the message box, options will show up of articles that the user can click on; those articles are then either shown within the "chat" window or in a new tab. There may also be buttons within the "chat" that can also be clicked.
Pros:
People have clear expectations of what they can get help with
Repurpose existing content on website
Means that there are persistent links that anyone can use to provide someone with the most up-to-date answer on a given question
Cons:
All answers have enough content to warrant their own page, and so questions with shorter answers may not be handled as well
Vendors:
MedChat
ServiceNow
Zendesk
We have only been able to find results about chat’s effectiveness at reducing call volume for states that have implemented “prompted” chat systems, and from only two of the vendors.
Colorado (Google Chat Bot): “the virtual agent solves user questions about 90% of the time”
Illinois (Google Chat Bot): “the web chatbot interacts with upwards of 100,000 constituents a day.” Google Chat Bot
New Jersey Google Chat Bot): “In its first 3 days of operation, the chat feature engaged with approximately 50,000 user interactions, freeing up time for agents to focus on claims that need intervention rather than answering frequently-asked questions.” Google Chat Bot.
Texas (AWS & Accenture): Over several months, answered Qs from 2.3 million people and save their staff from answering hundreds of thousands of calls
While all this data is promising, it leaves out how many people were confused by the chat system or just otherwise didn’t get their questions answered. That said, if your chat system is set up to help people get answers to the most common questions you see coming in through higher-cost support channels, chances are that you are well set up to deflect those inbounds.
If you are interested in working with USDR on updating your customer support mechanisms, don't hesitate to reach out by filling out this form.
How workforce agencies can make smart investments in digitizing employers’ experiences. Policy Memo from USDR, compiled by Alyssa Levitz based on research with Alissa Rubin and Emily Alter, June 2021
Employers play a key role in the ultimate disbursal of Unemployment Insurance funds; whenever someone files a new application for UI benefits, their former employer is notified and given the opportunity to challenge that claim. The data that employers contribute is assumed to be “more correct” than whatever the employee had originally provided and so has a lot of power in determining whether someone is in fact eligible for benefits, and if so, how much. How effectively employers interact with UI systems has a direct impact on the timeliness of payouts to legitimate claimants and could also indirectly reduce fraud.
In January and March 2021, we conducted 14 one-on-one, hour-long interviews with human resource professionals who handle unemployment claims for their companies. The goal of the interviews was to create a baseline understanding of their experiences in three key areas:
Submitting claims on behalf of their former employees when applicable
Being informed of claims filed by former employees and of any subsequent decisions made
Disputing and/or appealing decisions about a former employee’s claim and providing documentation/keeping up with the adjudication process
The sample was national, with participants based in 10 states: California, Georgia, Illinois, Massachusetts, New Mexico, New York, Pennsylvania, Texas, Virginia, and Washington State; however, most participants managed UI claims in multiple states. Interviewees worked for companies that ranged from dozens of employees to over 10,000 employees, in many different sectors: regional public transit, catering, health care, software development and more. Some of the companies experienced regular, seasonal layoffs that resulted in regular and predictable interactions with workforce agencies. A few of the companies whose representatives we interviewed were “reimbursable” employers, who were responsible for covering 100% of the benefits their former employees used; the rest of the employers were “taxable,” meaning that their payroll tax rate was determined by how heavily their former employees used the system.
Paper mail-based systems broke down when people were no longer regularly in the office
Delays occur in benefit disbursal when the claimant files for UI under a different name than the employer has on record, which happens much more frequently for people with non-English name patterns.
HR employees who worked at companies with a presence in multiple states either had to use a Third Party Administrator (like ADP or Equifax) or build their own internal processes to manage claims across the disparate systems.
Even in states where the workforce agency made online claim management possible, employers still encountered challenges in keeping up with new claims and providing the necessary information to respond to a claim.
There is no one way that employers think about their employees receiving UI benefits; some see it as a key part of their business model and thus do their best to help their former employees receive benefits, and on the other end of the spectrum, some employers have an institutionalized antagonistic relationship with former employees over receipt of UI.
Employers dispute initial claims, in particular surrounding reasons for “job separations,” but rarely dispute decisions
In the form or questionnaire provided to employers to respond to a claim, they should be given the ability to mark likely fraud and the space to provide more information about their perspective on a claimant’s situation.
Workforce agencies should all have a digital portal for all employers to at least learn about and respond to new UI claims, if not follow them through any appeals processes. This digital portal should:
Support multiple log-ins for a single employer, with easy and quick processes for creating that new account
Support mass uploads and downloads of many claimants’ information at a time
Let portal users opt into email notifications for new claims and newly scheduled hearings
Let portal users change or update their previously submitted answers to questions when they learn new information.
Increase transparency between employer and claimant so that simpler issues could be resolved more quickly
Have accessible, prompt, and employer-specific customer service options, for example:
Dedicated phone lines with staff knowledgeable about the employer experience, that do not solve claimant issues or even redirect claimants to the correct line (while extreme, some states have found this useful in keeping the phone line as clear as possible)
Creating FAQ and other online support content for employers to self-solve their issues
To see more detail on the findings and key recommendations, read the two research reports: Landscape of the Employer Experience (Jan. 2021) and Employer Experience: Digital Interactions (March 2021).
To see more about the way the current system of interacting with employers affects workforce agency processes staff, see the Employer Interaction section of the UI Journey Map.
Lead User Researcher: Alissa Rubin
We conducted interviews with 6 HR and payroll professions to learn about the way they interact with unemployment insurance agencies and systems; they collectively had experience with 15 states' systems.
All of them had to come up with their own ways to check their company's records with what the UI agency sent them, and they had difficulty at times in most states sending information back to the agency (the form did not meet their needs, or they were unsure if they were filling it out correctly). The amount of paper to track in non-digitized systems is astronomical, but the digitized systems also have their own challenges for responding to and tracking claims.
Learn how employers or their representatives interact with the UI system, including understanding employers’ level of knowledge around how their actions affect UI claims or payments, and what impact they aim for in those interactions
Understand the ways in which hiring and termination strategies are impacted by the UI system or vice versa
Based on the goals above, an interview guide was created. First, personal networks were leveraged to reach business owners and HR representatives in California for 3 initial interviews over the phone. This allowed us to test-run the interview guide.
Some additional questions were added to the interview guide at this point, before using the online recruiting tool UserInterviews.com to recruit appropriate participants and schedule further interviews over video conferencing. Interviews were scheduled for hour-long windows, although some were shorter. We used digital research tool Lookback.io to host and record the video sessions and enable screen sharing from participants. In some cases, Zoom was used rather than Lookback due to technical difficulties.
Gender
Race
State
Age
Company info
Participant 1
Male
White
Georgia
32
10,001+, taxable, not seasonal
Participant 2
Female
Hispanic
Massachusetts
40
200-1000, reimbursable, seasonal
Participant 3
Female
White
Illinois, but also deals with UI in: AZ, CA, IN, MA, OR, PA, TN, WA, VA & others
37
200-1000, taxable, seasonal
Participant 4
Female
White
New Mexico
39
200-1000, taxable, not seasonal
Participant 5
Female
White
California, but also deals with UI in MI, NC, WI
27
50-60, taxable, not seasonal
Participant 6
Female
Black
Pennsylvania
59
200-300, taxable, not seasonal
Dealing with the mail is frustrating and slow. Mail can get lost or sit for several days (or weeks during COVID) before getting delivered to the HR rep or being picked up from a closed office. This significantly reduces timeliness to respond to claims, and sometimes deadlines have already passed. HR reps also must manually compare these paper forms to (usually) digital employee records; they might scan all such claims forms for their records as well, and/or undertake other digital processes to keep track of paper claims, such as creating spreadsheets or emailing themselves digital copies of forms.
Challenges range from a frustrating user experience (e.g., a lack of notifications, difficulty navigating the portals to complete tasks) to major functionality issues (e.g., forms that won’t download or open; incompatible data and files that force manual data entry, which is error prone; inability to complete some steps on certain browsers).
Additionally, companies operating in many states had to deal with separate portals for each state, which all function differently and can be hard to manage.
Employees and Employers aren't inherently antagonists within the UI system: Participant 3's employer is already taxed at the highest rate for UI and actively provides advice and support to former employees to help them navigate the UI system; Participant 5 was told by their CEO that the employees should get benefits as much as possible.
On the other hand, Participant 1 was open about wanting to dispute claims, seeing it as his responsibility to the company, and Participant 6's company fires many employees for things that they believe are "cause" and ends up disputing many claims.
When asked, HR reps were interested in preventing fraud, but had few resources to do so other than maintaining solid payroll records and using these to compare accuracy of claims. Many relied on their personal knowledge of current and past employees to spot fraudulent claims. Several participants noted that they don't know how to mark a claim as fraudulent when they see it come through.
Participant 1 in GA hadn't gotten to the adjudication phase of any of the COVID-era contested claims at the time of the interview.
Participant 2's company in MA didn't receive a bill for 4 months last summer
“I decide what I contest, and I know what is and isn’t eligible. I don’t waste my time contesting claims I know I’m going to lose. ... and I’ve been right 100% of the time. It’s a huge error on my part to invest time if we aren’t going to win.” --Participant 1
“I don’t trust that DUA [in MA] is reading all the comments, and I feel like maybe something could slip through the cracks that would negatively impact an employee’s claim.” --Participant 2
“More than 95% [of this job] is just mail. Just paper, hundreds of pages. I’m not being sarcastic. Some of them are multiples of the same exact piece of paper coming in two different days, I’m trying to compare them. And then you have some states like AZ that’ll put three to four people in one envelope.” --Participant 3
“When I look at the website, it’s really hard to answer and get to my open claims. Filling them out with the paper is more effective as far as getting answers. Half the time I can’t see some of the claims that have been filed; I don’t think it’s me, I think it’s the website.” --Participant 4
“I would be grateful to have more resources from states on how to deal with fraudulent claims. I didn't even know how to fill out the form.” --Participant 5
By Emily Alter, User Experience Researcher & USDR Volunteer
Research conducted March - April 2021.
These recommendations are based on 14 hour-long interviews with human resource professionals who handle unemployment claims for their companies. The sample was national, with participants from 10 states: California, Georgia, Illinois, Massachusetts, New Mexico, New York, Pennsylvania, Texas, Virginia, and Washington State. Most participants managed UI claims in multiple states.
The first round of interviews, with 6 participants, aimed to understand employer needs and challenges with the UI system as well as their knowledge of the UI process.
The second round of interviews, with 8 participants, focused on employer needs and challenges in relation to the specific tasks involved in the administering UI claims. It also included a prioritization activity in which participants ranked the importance of each UI-related task for their company.
Participants were recruited via UserInterviews.com, a research participant recruitment tool.
The first round of research selected for employers that had laid someone off in the last 12 months.
The second round of research screened for HR professionals who had managed more than 5 unemployment claims. The participant selection process for the second round of research also favored HR professionals who had managed more than 25 unemployment claims prior to March 2020 and who had been involved in disputes or appeals of unemployment claims.
These screening criteria, especially for the second round, favored mid-sized or large employers.
This report describes people claiming UI benefits as claimants rather than former employees because they may be
Former employees entitled to UI benefits;
Former employees not entitled to UI benefits;
Current employees who went on furlow or had their hours cut, now returned to work but mistakenly or knowingly continue to receive UI benefits;
Former contractors not entitled to UI benefits, confused about their benefits; and/or
People committing fraud.
Overall, there is strong support for a digital portal. Right now, employers mostly receive notice that a claim has been made via paper mail.
A digital portal would decrease processing time. Paper mail is particularly difficult during COVID times because mail is slow to arrive at the office and the people in charge of UI claims are rarely at the office. Even outside of COVID times, once a letter reaches an office, the letter may take time to get to the person in charge of processing UI claims, which reduces the time the person has to respond.
Paper mail is inconsistent with modern business processes. Most business is now conducted digitally. Storage of employment records is now often digital, although many companies have paper backups. The state sending letters means that HR folks at employers need to take the time to scan them into their digital systems.
“In this day and age, to not have an electronic system seems crazy to me.” -California reimbursed employer, 1000+ employees, received claims via a TPA that shared claims as a spreadsheet
“We get them all by mail. It goes to the front desk, who takes it to payroll, who scans it. By the time I have to make the response, I’ve just got it because of all the passing around.” -Washington State taxable employer, 1000+ employees
“It would be so much more efficient for all parties if it was sent electronically.” -California taxable employer, 100+ employees
The portal needs to support the UI process for a range of businesses. Employers who participated in the interviews were either small companies that handled their UI processes themselves or larger companies that used a TPA (third party administrator) to handle their processes in bulk.
Learning about new filings was the most important UI task for the majority of employers, with the exception where a TPA handled new filings for them. Learning about new filings was the most frequent UI task. Usually companies had 2 weeks to respond. If they didn’t respond in that time, then they might start getting charged for UI.
Disputing claims and satisfying requests for additional information were the most important UI task for large organizations whose TPAs handled new filings and the second most important task for other organizations.
Most HR professionals had learned through experience which claims they could dispute and win in different states. They prided themselves on only disputing claims they could win. However, a few worked for companies with a policy of disputing all claims.
Satisfying requests for additional information from the state was also one of the most common tasks related to unemployment insurance. HR people wanted to provide that information to the state promptly.
Support employers in accessing information in a timely and secure fashion. Support timely new employer account creation. Organize portals via individual claimant as well as employer. Support multiple log-ins for different HR employees at a single employer.
Support timely new account creation. Right now, creating and verifying a new employer account with a new state can take more than 2 weeks, which is longer than employers may have to respond to a new claim. (Many employers have employees in multiple states.)
“I went to do the registration on the California website [for their only California employee], and I found it very difficult to manage. It says, after two weeks, we'll add you into the platform. I ended up handwriting everything and mailing it in instead of going on the portal. With most states, it's not an easy process to just log on to the website, find your company and find everything to manage.” -Virginia state, 10+ employees, 200+ contractors
Digital portals should be organized by employer and support mass upload and download. Consider also allowing employers to sign in with a unique password and pin to access individual claims. Most digital UI portals are already organized by claimant inside the employer account. This makes sense to employers.
Support mass uploads and downloads for employers. Digital portals should support mass downloads and uploads of information related to UI claims. Larger companies, TPAs, and companies with seasonal businesses and regular or unexpected fluctuations in their workforce size need to be able to handle claims in aggregate. Forcing employers to individually process data for tens, hundreds or thousands of claimants at a time will result in a slower process for claimants.
A Pennsylvania taxable employer with 1000+ employees that used a TPA for most of its operations acquired a new company that did UI claims manually. “It's the same information, but it just takes a lot longer.. to fill out all the information, attach all of the documentation manually, and then copy them all, scan them, and then fax. It's a hassle.”
Support multiple log-ins for a single employer. A log-in for an employer being tied to the personal data of one or two people at an employer is difficult for employers when multiple people at an employer handle UI claims. Support a single employer having multiple log-ins that different individuals at the employer can use.
“We can only have one or two usernames for the portals, and it’s not good practice to share usernames. Even just payroll has 5-6 people. There’s no way to have everyone to log in individually.” -Washington State taxable employer, 1000+ employees, where responsibility for managing UI claims is shared between payroll, the HR person who handles union members and the HR person who handles staff.
“We might need me, the COO/CFO, and CEO to access the (company) account.” -California taxable employer
Consider allowing employers to sign in with a unique password and pin to access an individual claim, without having to create a new account for themselves as an employer. The password and pin should come with the notification that they have received a new claim. This will speed claim processing time by not requiring the creation of a new employer account.
Consider collaborating across state lines to create a portal that can aggregate and simplify processes regardless of state. Most of the HR professionals in this research managed unemployment for their company in multiple states (see “About the research” at the end of this report for how the participant selection process favored mid-size and large employers). They disliked having to learn a new user interface for each new portal. It made them less efficient at processing claims. Having a portal that worked for multiple states would reduce the number of new digital portals that multi-state employers need to learn.
Notify employers about new claims via email and portal. Default to email, but allow companies to specify that they would prefer notifications about new claims to come via portal.
The most convenient notifications from outside a company come via email. These days, emails are a standard form of communication for most external notifications from outside the company. Emails have the advantage that they support asynchronous communication. Someone can get an email and take a few hours to find the requested information and compose an answer. In contrast, someone has to be there at the time to answer the phone. And the person who picks up a phone call may not have the answers at their fingertips.
Some companies and TPAs support a large enough volume of claims that they prefer a portal. For large employers and TPAs, a portal is more convenient than pages of individual emails. One HR person who used a TPA described going into the TPA portal to check on UI claims as part of their morning routine alongside getting coffee and checking her email.
“I would want a file feed to go directly to the benefits portal from payroll, like all our other systems.” -New Mexico taxable employer, 200+ employees
Confirming the details will benefit from some transparency between the employer and claimant. In most states, employers find out about a claim when they receive a letter from the state saying they have a claimant. Then, it is up to the employer to confirm various details, such as the employee’s name, their dates of employment, and the cause of termination.
Suggest transparency between the employer and claimant on easily resolvable details. Easily resolvable details such as claimant name and their end of employment date should be shared between the employer and claimant to speed up claims. The claimant may also not be clear on the employer’s identity when multiple organizations are involved in their employment (ex., as a contractor).
The end of employment date given by the employer is sometimes different from the day of termination provided by the claimant (i.e., one fills out the form with the last day they showed up to work and the other fills out the form with the day after). This can slow down processing. Where employers and claimants have the same interest in speedy processing, transparency on the end of employment date in the best interest of both parties.
The employer’s identity may not be clear to the claimant. For example, employees at a cafeteria on a college campus may not understand they were working for the company that provided catering, not the university itself.
Consider transparency between the employer and claimant on more contentious issues.
“I don’t want this responsibility, but it would be so much easier if I could see what employees are putting in on their UI or talk on their behalf on some point. If you make a call to the cable company, you can do it if you’re verified on the account. I can help interpret because it never comes out in English.” -Illinois taxable employer, 200+ employees
The cause of the employee leaving the company. When they file, the claimant and the employer may provide different reasons the claimant is no longer employed by the company. Sometimes this represents a legitimate disagreement. Other times, the claimant may not understand the categories offered in the form and/or may not be completely fluent in English. If the employer and the claimant are in agreement about the cause of the employee leaving the company, transparency about the other party’s claims may reduce the need for further action.
“A lot of times, not usually on purpose, people get confused and report a different reason. People have been awarded UI when they weren’t supposed to. They go to fill out the form and they’re not sure the difference between layoff and fired.” -Washington State taxable employer, 1000+ employees
The nature of employment. In a world with contractors and agencies, claimants may not be clear on the nature of their employment (ex., a 1099 vs. a W-2). For example, a contractor may claim unemployment benefits to which they are not legally entitled if they were on a 1099.
Accommodate complexity. The circumstances and people involved in an employee leaving a company can be complex. It is useful to accommodate that complexity by providing an “other” option in the reasons for the end of employment and a place for notes. It is also useful to allow employers to change their answers when they learn new information.
"We had someone that was a really volatile person... I knew the hearing was going to be contentious. It was a six hour hearing. That has never happened to me. They're usually under an hour.” -Pennsylvania, 1000+ employees, used a TPA
Provide an “Other” option in reasons for the end of employment and provide a place for notes. Having a place for notes in forms accommodates the complexity of some of these situations.
“Not every situation fits into a questionnaire neatly, and there’s rarely room to add a full story.” -PA taxable employer, 10k+ employees, used a TPA
Allow employers to change or update their previously submitted answers to questions when they learn new information. HR departments may learn new information in the aftermath of an employee’s departure that changes the former employee’s UI eligibility (ex., evidence of stealing on the job). In the case of a senior leader departing a company, there may also be ongoing negotiations between the senior leader and the employer which changes information that may have been initially supplied to the state.
“Sometimes things change. Someone will put in their notice and then get fired, so how do I change that?” -Washington State taxable employer, 1000+ employees.
Design to accommodate data from multiple sources which may be in incompatible formats. Support the combination of multiple records into a single record, and the separation of a single record into multiple records. For example, a father and son may have the same name and address. Design for the flexibility to separate a single record into multiple records if their information was incorrectly combined into a single record.
Support claimants having different names. The system should support the combination of two or more records if the same person appears in UI systems under different names. Sometimes the same person will appear in UI systems under different names, whether that is due to system limitations (i.e., character limits not supporting a long last name), cultural factors (i.e., name changes upon marriage), or human error (a middle name appearing in one form and not another).
“If Jake’s name is really John and he put Jake on his application, he has to update all that, so I have no idea if they’re putting in the right information.” -Illinois taxable employer, 200+ employees
Don’t limit name length. Support diverse names, including both shorter (last names with only two characters, like Li) and longer names. Either have very generous character limits or don’t have character limits. Support apostrophes, hyphens, spaces and special characters (ex. Juan Fernández de Calderón García-Iglesias).
“Certain communities have names that are longer. Hispanic communities, it was a big problem. Names would have a character limit and they would have issues with their claims because of that. The system is not robust enough.” - Georgia taxable company with 10,000+ employees
Support name changes. Sometimes, people change their names (ex., upon marriage). Systems should support name changes and allow two separate records under different names to be combined.
Lack of flexibility in the scheduling of hearings could be a minor frustration for employers. Preparing for hearings usually required research into the history of the case. In most states, employers were not given a choice of when the hearing would be. If they received notice of a hearing with only a few days to prepare, they had to drop other things to focus on preparing for the hearing. Also, one participant reported having to wake up at 5 am one day to attend a hearing in another time zone. Suggest requiring that employers receive at least a week’s notice of a hearing and slightly more flexibility in the scheduling process.
Employers disputed claims, but rarely disputed decisions. The majority only disputed claims they thought they could win, although some had a policy of disputing all claims. The portal user interface can visually emphasize other options than the option to dispute decisions because it is a low frequency occurrence. HR professionals learned through experience that they would rarely win if they disputed a decision, so they disputed decisions infrequently if ever.
States should provide clear options for employers to report UI fraud. They should also consider pro-actively auditing for UI fraud.
Provide a clear place in the forms for employers to report fraudulent claims. Many states don’t currently provide an easy way for employers to report fraud, ex., if they see a claim from someone who has never worked for them or is currently actively employed with them.
“Especially with COVID, there was a huge influx of fraudulent claims that came through, and there is nothing here on this screen that something is fraudulent, so for the first month of COVID, I had to call every time I thought something was fraud.” -Massachusetts taxable employer, 200+ employees
Consider pro-actively auditing for fraud. None of the employers pro-actively audited for UI fraud until it came across their desks. After it came across their desks, they were aware. But if it hadn’t come across their desks, they didn’t check for it.
“The other thing the TPA did during the pandemic that was very helpful was tipping me to places where they thought fraud had occurred.” -California reimbursed employer, 1000+ employees
Have accessible and prompt customer service. Some HR folks did not know how to contact the state if necessary. Others dreaded waiting on hold for extended wait times. Accessible and prompt customer service for employers ensures that states can process claims efficiently and get them out to qualifying claimants. Additionally, sometimes former employees will contact HR for help in getting through extended wait times. Employers couldn’t help and it increased their frustration with state UI systems.
“You should be able to pick up a phone.. It was a lot of red tape from this person to this person (at the state unemployment offices)” -New York State professional, discussing the unemployment claims process at their previous 1000+ employee company
Unemployment insurance (UI) is a deeply complex political, legal, social, and technological space. In order to try to make sense of it, we have created a journey map to visualize not only how people interact with the system but how the system itself works, including what UI agencies have to do to administer it. UI systems have been underfunded for a long time, with challenges and vulnerabilities that were only exacerbated by the unprecedented volume of claims during the pandemic.
This report summarizes some of those challenges within the journey map framework, and provides recommendations for how states, the US Department of Labor, and other key actors could increase the effectiveness and efficiency of UI administration.
Some states require you to certify within certain time windows (as small as 30 minutes) as a way to balance web traffic load.
It increases the amount of times that someone is going to be able to mess up their claim: "It’s like running the gauntlet every week. If anything changes, one error will throw you down a path that will cause delays for the claimant and us that may require manual intervention Weekly certification is something that US DOL will not relax even though everyone has asked, maybe once every two weeks, because of sake of anti-fraud measures, they see this as another checks and balances in regards to certifying for benefits." - State B
Agencies should find alternative ways to balance traffic loads. The load generated by certification should be extremely small.
Certification questions are complex and difficult to understand and updating them requires the state to work with the Department of Labor. People are nervous that if they click the wrong option, they’ll lose their benefits.
"We identified questions that were confusing to claimants. Available to work and telework questions...the original language that DOL recommended and drafted was very confusing.” -- State C
"Systems haven’t been translated to serve underrepresented populations [or people] who don’t speak “UI language,” which is not human language, but it’s built into our system. Some of this language is mandated by DOL; it’s been adopted as the norm across policy and programs.” -- State F
US DOL should provide guidance and draft communications for states that has already been put into into plain language and usability tested for clarity in English, Spanish, and other key languages as requested by states.
If US DOL or one of its regional offices works with an individual state on changing any communications, that should be shared out with other states so everyone can benefit from that and doesn’t need to be re-done by others.
Journey maps are a way to visualize how users interact with a system. In Unemployment Insurance, there are four key user groups, each represented by a different color on the map:
People seeking UI benefits, “claimants” (light blue)
Government staff administering the program, “agency” (gray)
Companies that pay into the system, “employers” (gold)
Other government actors: Federal and State legislatures, Governors, and the US Department of Labor (dark blue)
Each box of the journey map represents a key step in the process. The unemployment insurance (UI) journey begins in the top left with a claimant trying to figure out whether they are eligible for any UI programs at all.
As claimants go through the process from left to right, the UI agency does a lot of work in the background. The agency takes direction from US Department of Labor as well as state-level stakeholders like State Legislatures and Governors; they also report back to those entities as directed. The agency also facilitates communication with employers where appropriate to help validate information provided by the claimant.
This map is not intended to be a complete picture of the UI system; for example, it doesn’t capture what happens if someone is found ineligible at any point. As our research continues, we expect to refine and expand this map over time.
Unemployment insurance is a complex system with many factors impacting eligibility and compensation amount. The way that UI agencies determine and communicate monetary and non-monetary eligibility independently is confusing: e.g., you’re told that your situation may qualify you for UI benefits, but also that you didn’t work enough hours in the base period to actually get any benefits -- or you’re told that you would get a certain amount if you qualify for UI benefits but that you don’t actually qualify.
The complication of UI systems was exacerbated when the 2020 CARES Act added three new programs to be administered through UI systems: Pandemic Unemployment Assistance (PUA), providing a base of benefits for gig workers and the self employed who are not eligible for standard UI; Pandemic Emergency Unemployment Compensation (PEUC), which extended the number of weeks someone could collect standard UI; and Federal Pandemic Unemployment Compensation (FPUC), which initially added an extra $600 to both UI and PUA for a period of time.
People were getting information about these programs from the news, from their friends and family, and from various forums online in addition to what was made available by the administering UI agency; the lack of straightforward explanations led to incorrect expectations about what they are eligible for and how much they would receive.
"It’s like an evil MC Escher painting. You think you’re walking through one door of eligibility and end up on the stairs of something else" --State B
"There are so many programs [in addition to unemployment insurance]: disability, temporary leave, etc. It’s hard for a consumer to know [what is going to be the best fit for them] unless they go through fantastically designed charts or FAQs, which they don’t usually do" -- State B
"Just straight UI claims - people just say give me my money, [but] there are lots of eligibility requirements around both monetary [amount and situations] with rules and laws. It’s not just about the fact that you’re not working,” and that’s hard for people to understand. -- State B
Agencies should se plain language when explaining programs and eligibility.
When there are multiple programs that someone might be eligible for, agencies should ideally implement a benefits eligibility screening tool, but at the very least provide an eligibility checklist, or eligibility FAQ page to help guide claimants to the right program(s).
Agencies should make it clear that the benefit is taxable and provide an opportunity to withhold that amount (if state law doesn’t permit withholding, that should change in partnership with the state legislature). If the claimant chooses not to withhold, the implications of not doing so should be clear in the application.
Agencies should perform usability studies on how people are informed of their (in)eligibility determinations to understand what specific communications should be streamlined and/or re-written. Communicate benefits receipt as a clear approve/deny in a single letter, with additional information about benefit amount or reason(s) for not qualifying for any benefits.
To be eligible for Pandemic Unemployment Assistance (PUA), a claimant must first apply for standard unemployment insurance and be rejected before applying for PUA, many times with the same information as in the initial application. States did not do an effective job communicating this to claimants and helping them navigate the process.
The impact is an unknown decrease in benefits participation rate, which resulted in people not getting the help they need and are qualified for.
"There is a DOL requirement that we deny people UI before we give them PUA. So we have people understanding that they can apply for PUA, should apply for PUA, but why they were denied for UI is super confusing. [We should have been] able to do some assessment and offer you just the PUA link, that’s what you’re eligible for based on your application. [...] It’s amplified the risk that people don’t understand what’s happening. [...] There’s a pretty big drop-out rate there where they don’t then go to apply to PUA, but we haven’t spent time hunting that down." -- State F
Agencies should connect applications when possible: If there are programs whose eligibility depends on not qualifying for other programs, application information should be sent from one to the other so that claimants don’t need to re-enter information they’ve already provided.
If it is going to take any amount of time to determine whether they are eligible for the first program, agencies should communicate this clearly and let them know what to expect.
If someone is found to be ineligible for the first program, agencies should use the information already obtained to determine whether they might be eligible for the second program and help them through that process. Do not rely on them knowing that they need to apply for something else.
New programs take time to establish and they can be difficult for state agencies to set up. People get confused and angry about the delays, believing that benefits should be available as soon as the program is mentioned in media coverage.
Legislation doesn’t take into account the feasibility or ease of enacting programs: “Every time Congress sneezes, we catch a cold. We went through this during the last recession, too. Let’s figure out what this should look like next time; assistance should be built atop existing code that can just be activated" -- State C
"What people hear is that there’ll be $300 more in your account tomorrow. And that’s not how it works, we have to reprogram everything." -- State D
US DOL should consult with states when any federal timelines are established to ensure that there is a reasonable implementation period. US DOL should help states in communicating the implementation timeline, especially if the legislation did not provide sufficient time.
Agencies should be as clear as possible about when the program will be operational; communicate changes to that timeline as they happen. Remain empathetic about what the claimants might be experiencing as they await aid.
US DOL should engage with states in a user-centered policy making process to understand the challenges with implementing new programs during crisis. That information should then be used to inform emergency legislation in future crises, as well as grant programs for UI agencies to prepare for recession-era programs ahead of time.
Phone lines make it hard for agencies to prioritize work in a strategic way; you have to work with whoever calls in rather than, e.g., working on the oldest applications first.
Any published phone number is going to get inundated with calls, even if it’s intended to serve a particular purpose, because most states’ systems don’t provide people with enough information online.
Phones are a very expensive way for people to get the help they need
"People can’t self-cure almost anything on their application, even when they realize they’ve made a mistake, they can’t do anything about it until they talk to an agent" -- State B
"You can look up generalities, how to file, what to do, but once the claim is filed, someone generally wants claim specific, and the only way we can provide that is by speaking to someone, and we’re way outnumbered. The ability to get to a live agent, sure it’s a problem." -- State B
People who might actually need the phone (i.e., those without proficiency in or access to the digital tools) can’t access them because “we are clogging the lines with UX issues because it’s not clear or understandable to anyone.” -- State F
"There are escalation queues for people with disabilities, for people whose primary language isn’t English, people with significant hardships, etc. But if they advertise any of those routes, calls will come in and fill that up. Specialty phone lines don’t work." -- State F
US DOL should provide states with plain language, usability-tested content (graphics, FAQs, descriptions, social media posts, etc.) of the concepts that are the same across states, especially: how the different UI programs relate to each other, how to understand eligibility, what the impact of responding to weekly certification questions is, etc.
Agencies should implement chatbots, eligibility screeners, callback scheduling systems, and other tools to reduce the need for people to call in.
Agencies should have clear processes in place to regularly learn from its customer support mechanisms (FAQ traffic, chatbot interactions, call center topics) and leverage that to update website and chatbot content.
Agencies should conduct usability testing of the key parts of their experiences to identify opportunities for improvement so that people don’t need to call in order to understand the system or find out what’s going on with their claim.
To learn more about all of U.S. Digital Response's work, see our website. Get in touch with us by filling out our contact form.
This guide can be found on GitBook and on GitHub.
Many of the limitations of unemployment insurance systems existed before the pandemic hit; the difference is that before the pandemic, unemployment claims were predictable enough that agencies could appropriately staff up to account for the constraints. When COVID-19 began to spread and businesses started en masse reducing work hours, conducting layoffs, or closing all together, unforeseen and unprecedented numbers of people were filing for unemployment insurance benefits. Because the UI systems are set up to scale with the hiring and firing of individuals, many states are still months behind in processing claims.
The human toll of these backlogs is real. Without the UI benefits delivered in a timely manner, many more people have had to make impossible financial decisions: food or rent? Utilities bill or medicine?
The Unemployment Insurance Modernization team at USDR is working to understand the constraints under which UI systems operate and, by partnering with those agencies, evaluate, plan, and/or implement solutions to help them become more effective.
A non-comprehensive list of the articles, reports, and publications that have been useful in developing our understanding of UI systems in the United States. Originally compiled by Alyssa Levitz.
"Pandemic Unemployment and the Social Safety Net" funded by New America:
"Barriers and pain points of applying and getting assistance" by Lindsey Wagner & Dana Chisnell
"Time spent applying and waiting for benefits" by Gina Kim, Lindsey Wagner, Rebecca Michelson, Grace Cho, & Dana Chisnell
"The Unemployment Pandemic: Addressing America's Job Crisis" by Michele Evermore
"Unpacking Inequities in Unemployment Insurance" from New America, by Monée Fields-White, Vivian Graubard, Alberto Rodríguez, Nikki Zeichner, & Cassandra Robertson
2020-10-05 The Century Foundation, National Employment Law Project, and Philadelphia Legal Assistance: Centering Workers: How to Modernize Unemployment Insurance Technology.
This report provides an analysis of states' pre-COVID efforts to modernize their UI systems in order to make recommendations on how to continue this effort with more success moving forward.
The recommendations they make are even more important to consider as COVID-era volumes underscore the existing limitations of UI systems; the recs are broken down based on stages: Planning, Designing, and Implementing.
Within the "The Role of Unemployment Insurance" section, they do the important work of sharing some of the ways in which systemic racism impacts both UI systems and the underlying unemployment that make the systems so critical. This is a fundamental piece of their argument for the inclusion of the affected populations in making major changes to the UI systems.
From Disrepair to Transformation: How to Revive UI IT and Infrastructure, Congressional Testimony from Rebecca Nixon, Executive Director, NELP
2020-06-08 NASWA, Upjohn Institute for Employment Research Study: Implementation of the American Recovery and Reinvestment Act: Workforce Development and Unemployment Insurance Provisions. (Also see: report pdf).
This report contains key lessons learned from the implementation of the American Recovery and Reinvestment Act (ARRA) enacted in 2009.
From National Association of State Chief Information Officers (NASCIO): Chat with Us: How States are Using Chatbots to Respond to the Demands of COVID-19. This report isn't exclusively about unemployment insurance, but it includes information about UI chatbots.
2020-11-30 -- Government Accountability Office: GAO-21-191: Covid-19: Urgent Actions Needed to Better Ensure an Effective Federal Response. There were 4 key take-aways for UI in this report:
The Federal Department of Labor had been reporting the number of weeks of UI benefits claimed as the number of individuals claiming benefits. Given backlogs and multi-week certification processes, this has the effect of inflating how many people appear to be on unemployment.
UI Benefits are funded by taxes levied by states on businesses; these funds are completely used up in many cases. “As of November 9, 2020, 21 states and territories held about $40.2 billion in federal loans to pay UI benefits.” It will take years after this recession to re-establish the funds and pay back these loans.
Most states have been paying the minimum allowable benefit for PUA rather than calculating what folks are truly eligible for. The DOL doesn't know how many states are tracking, recalculating, and disbursing these payments, though California appeared to be doing so as of October 30, 2020
There are a lot of on-going fraud investigations. 14 investigations by DOL itself, nearly 150 by the US Secret Service, and unknown numbers through money given by the federal DOL to states.
2020-09-16 -- California Employment Development Department Strike Team: Detailed Assessment and Recommendations
This 100+ page report with appendices details the innovations and limitations within CA's EDD
Its primary recommendations were:
Measure and monitor every step of the claims process daily
Deploy a NIST IAL2-compliant identity proofing solution within the application process
For a period of time, delay new claims filed and redirect highest skilled staff to focus on processing claims
Complete deployment of mobile-friendly document upload solution
2021-01-28 -- Nevada DETR Rapid Response Strike Force Report****
“How DOL reaches out and the language they use matters. Right now, they only reach out for audits, never see them reach out like NASWA. Having DOL go on NASWA calls will answer a lot of questions. -- State A
“[Regional offices have] a lack of people that ever worked in a state UI agency, and I appreciate their empathy and support; I think a huge percent have never worked on a claim before, have never taken a call about a claim before, they don’t know what we’re talking about, they don’t know the process for a UI claim, they never really had to deal with it until this year." --State B
"They’ve become like auditors full time at workforce and UI. they don’t have time for anything other than making sure paperwork is correct and we’re meeting metrics. They don’t have time for counseling, outside the box thinking" -- State B
“It’s a gamble as a director, do I want to wait for the guidance, or hope I interpret it right? I don’t care where the guidance comes from, I’m not having to go to my governor and say we wasted X Million on doing this the wrong way. It feels like a gamble too often, paying people or waiting to figure out the rules. That’s a bad position all the way around." -- State E
"NASWA has done a good job on trying to share info and played a leadership role. What’s good is they are independent of DOL but also composed of people that have leaders who have worked in states and know what it’s like to be in our seat and be helpful and relieve burden. They’ve invested a lot.” --State F
US DOL should establish a culture of collaboration with states that reframes existing enforcement relationship
Meet with leadership in every state to demonstrate this
Provide air cover to state employees to contribute to new and innovative ways of solving challenges -- but also just to do their jobs
Work with states on new process to issue guidance iteratively and with input from states
Connect states with each other when common problems arise, and work with them to come to solutions.
US DOL should update reporting requirements from states to require the submission of a detailed breakdown and calculation of their backlogs to include reasons for entering manual review queues and the effectiveness of the manual review queues in resolving issues
Ensure states understand/have the technology support they need to report accurately number of individual claimants
Apply best practices (e.g., California report) to develop a national UI dashboard
As part of U.S. Digital Response’s pro-bono work to expand government’s capacity to deliver digital services, we offer language access services for state unemployment insurance agencies.
We can help optimize your unemployment insurance system through technical assistance with:
Translations grounded in plain language best practices and policy compliance to improve claimant comprehension
User research with workers in your state to inform data-driven improvements
Consultation on translation tech and workflow to scale multilingual support
These services aim to:
Reduce call center volume by preventing confusion for workers who speak a primary language other than English.
Lower administrative costs by eliminating errors and adjudications.
Increase trust and integrity by improving user experience across languages.
Contact us if you’re interested in language access support for your state.
Email feedback and suggestions for improving the translation guide to:
This is a plain language and translation toolkit with three focus areas:
The section on Writing Inclusive Spanish highlights writing tactics grounded in bias-free language.
The section on Writing Plain Spanish includes tips on translating plain English to plain Spanish.
The section on Designing Reader-Friendly Content for Spanish Speakers provides best practices for visually organizing information on a screen.
We’ve also produced a Plain Spanish Glossary of Unemployment Insurance Terms to use as a quick reference or as a data source for machine learning translation algorithms.
Visit Spanish Language Style Guides and Glossaries on Digital.gov to find more resources for improving how the government communicates with the public in Spanish.
In preparing the recommendations in this guide, we considered what constitutes an appropriate voice and tone for UI communications.
“Voice” conveys your agency’s personality and values. We recommend using a voice that conveys empathy, inclusivity, and respect when communicating with Spanish-speaking audiences.
Empathetic: Acknowledge that this can be a stressful time. Use a caring and trauma-informed tone.
Inclusive: Welcome all readers equally. Avoid assumptions and use neutral terms like "people" and "workers." Also, restrain from using gendered pronouns and age-related descriptions.
Respectful: Aim to be authoritative yet friendly. Be kind and use polite language. Make sure that the information and resources you provide are accurate, reliable, and consistent to build trust.
Adapt your "tone" depending on the situation, topic, medium, and audience. Key elements for an appropriate tone for UI may include:
Educational: Most readers are new to this information. Use examples to explain complex terminology.
Friendly: Use a conversational tone and welcoming words.
Instructive: Be clear and direct – get to the point and use action-oriented language. Draft content that is easy to understand at an 8th-grade reading level.
Motivational: Explain the "why" behind instructions to encourage action.
Updated September 29, 2023
This is a working list of plain Spanish translations and recommended usage for common unemployment insurance terms. All content contained in this glossary has been tested and validated for readability and comprehension with Spanish-speaking workers in New Jersey. We will be adding more terms to this list as we conduct more testing and continue to validate additional terms for plain language.
As of March 31, 2023, this list follows guidance from the U.S. Department of Labor's UI Lexicon Project and was developed from research conducted by U.S. Digital Response in collaboration with the New Jersey Department of Labor, Nava PBC, Truss, Legal Services of NJ, and workers in New Jersey.
If you are working on plain language or language access for unemployment insurance and are interested in receiving support in this area, we are here to help! Learn more about our Language Access work, or contact us anytime.
You can also request direct access to the spreadsheet or direct feedback and questions to language-access@usdigitalresponse.org.
Form intro content plus all 19 eligibility questions, each accompanied by legally compliant helper text optimized to increase readability, comprehension, and accuracy of responses.
View all Retroactive PUA Eligibility materials as Google Documents and download in your preferred format.
Issue: [INSERT ISSUE NAME]
Slug: PUA ampliación de elegibilidad (2020)
Title: Antecedentes
Nos ponemos en contacto con usted porque reconocemos que es una persona a la que anteriormente se le denegó la elegibilidad para los beneficios federales por desempleo debido a la pandemia, pero que ahora puede ser elegible para los pagos en base a las disposiciones de elegibilidad extendidas del último año.
Estas disposiciones se dan con efecto retroactivo desde el inicio del programa de beneficios federales por desempleo debido a la pandemia. Por lo tanto, debe responder a las siguientes preguntas para establecer su elegibilidad, aunque las haya respondido anteriormente. También reconocerá las preguntas anteriores que se encuentran en los formularios de solicitud semanal. Confirme todas las semanas y asegúrese de responder a las preguntas 17 a 19 relacionadas con las disposiciones de elegibilidad más recientes que abarcan:
Las personas que se niegan a volver a un trabajo inseguro o a aceptar una oferta nueva de trabajo que es inseguro;
Ciertos individuos que prestan servicios a instituciones educativas o agencias de servicios educativos;
Individuos que hayan tenido una reducción de horas o un despido temporal o permanente.
Nota: las personas que no presentaron una solicitud inicial para los beneficios federales por desempleo debido a la pandemia en o antes del 27 de diciembre de 2020 están limitadas a las semanas de desempleo que comienzan en o después del 6 de diciembre de 2020.
A menos que sea absolutamente necesario, por favor no nos llame con preguntas. La mayoría de las respuestas a sus preguntas se encuentran en nuestra página web. Usted debería encargarse de completar este formulario por si mismo.
Note: Translations for official questions 1-16 are not provided by USDR
1. Did your place of employment close as a direct result of the COVID-19 public health emergency?
Cuando contestar que sí:
su lugar de trabajo ha cerrado debido a la emergencia de salud pública COVID-19
no puede volver al horario habitual de trabajo debido a la declaración de emergencia COVID-19 o a los protocolos de distanciamiento social
Cuando contestar que no:
su lugar de trabajo no cerró en ningún momento durante la emergencia de salud pública COVID-19
su lugar de trabajo cerró pero puede volver a su horario habitual durante esta semana
su lugar de trabajo cerró por razones no relacionadas con la declaración de emergencia COVID-19 o los protocolos de distanciamiento social
2. Are you an independent contractor or self-employed individual, and has the COVID-19 public health emergency severely limited your ability to do your customary work activities?
Cuando contestar que sí:
es un contratista independiente o un trabajador autónomo Y ADEMÁS tuvo una reducción de su trabajo habitual hasta el punto de verse obligado a suspender actividades durante esta semana
Cuando contestar que no:
no es un contratista independiente o un trabajador autónomo
es contratista independiente o trabajador autónomo y ha podido trabajar su número habitual de horas esta semana
es un contratista independiente o un trabajador autónomo y puede realizar sus actividades laborales habituales (la actividad no es significativamente más lenta de lo normal)
3. Are you unable to work from home or anywhere because you are the main caregiver to a child or household member who needs your constant and ongoing care; and they cannot attend school, daycare or other facility that is closed as a direct result of the COVID-19 public health emergency?
Cuando contestar que sí:
no puede trabajar porque necesita proporcionar cuidados continuos y constantes a un niño o a un hogar
no recibe las prestaciones de permiso remunerado en virtud de la Ley de Respuestas al Coronavirus de Families First (FFCRA)
la escuela, la guardería o cualquier otro centro está cerrado debido a la emergencia de salud pública COVID-19 (se considera que la escuela está cerrada aunque esté totalmente en línea y no ofrezca clases presenciales, o bien ofrezca una mezcla de clases en línea y presenciales con alumnos que asisten algunos días a distancia)
Cuando contestar que no:
puede trabajar desde casa mientras es el principal cuidador de un niño o miembro del hogar
las escuelas, guarderías u otras instalaciones están abiertas
ha optado por permitir que el niño asista a clase a distancia (se considera que la escuela está abierta cuando se requiere la presencia física de los estudiantes o cuando los estudiantes tienen la opción de quedarse en casa y asistir a clases en línea o a la escuela en persona a tiempo completo)
4. Are you unable to reach your place of employment because of a quarantine imposed by a state or municipal order restricting travel as a result of the COVID-19 public health emergency?
Cuando contestar que sí:
su condado restringe los movimientos y, por lo tanto, no puede trabajar
Cuando contestar que no:
su condado no está en una fase de restricción de viajes
5. Are you unable to reach your place of work because you have been advised by a healthcare provider to self-quarantine due to concerns related to COVID-19?
Cuando contestar que sí:
se le recomendó la autocuarentena durante esta semana debido a la COVID-19
Cuando contestar que no:
no se le ha aconsejado la autocuarentena durante esta semana debido a la COVID-19
6. Were you scheduled to start a job in Washington state but no longer have the job or are unable to reach the job as a direct result of the COVID-19 public health emergency?
Cuando contestar que sí:
estaba contratado y programado para comenzar un trabajo, pero ya no tiene o no puede llegar a ese trabajo debido a la emergencia de salud pública de COVID-19
Cuando contestar que no:
no fue contratado y programado para comenzar un trabajo
está contratado y puede llegar a su sitio de trabajo
7. Did you have to quit as a direct result of COVID-19?
Cuando contestar que sí:
has dejado tu trabajo como consecuencia directa de COVID-19
Cuando contestar que no:
no ha tenido que renunciar a su trabajo a causa de la COVID-19
ha renunciado a su trabajo debido a la COVID-19 pero puede retomar su trabajo esta semana
8. Were you diagnosed with COVID-19?
Cuando contestar que sí:
ha sido diagnosticado con COVID-19 y sigue sin poder volver al trabajo (un diagnóstico calificativo no requiere una prueba positiva. Basta con cualquier diagnóstico de un profesional médico autorizado, incluido un diagnóstico realizado por teléfono o por telemedicina).
Cuando contestar que no:
ha sido diagnosticado con COVID-19 pero ha podido volver al trabajo durante esta semana
no ha sido diagnosticado con COVID-19
9. Do you have symptoms of COVID-19 and are seeking a medical diagnosis?
Cuando contestar que sí:
tenía síntomas de COVID-19, buscaba un diagnóstico médico y no podía volver al trabajo por este motivo
Cuando contestar que no:
no tenía síntomas de COVID-19
tenía síntomas de COVID-19 pero no buscó un diagnóstico médico
tuvo síntomas de COVID-19, buscó un diagnóstico médico y pudo volver al trabajo porque su diagnóstico fue negativo
10. Has an individual in your household been diagnosed with COVID-19?
Cuando contestar que sí:
alguien en su hogar ha sido diagnosticado con COVID-19 y usted no puede regresar al trabajo esta semana por esta razón
Cuando contestar que no:
nadie en su hogar ha sido diagnosticado con COVID-19
11. Are you providing constant and ongoing care for a family member or household member who has been diagnosed with COVID-19?
Cuando contestar que sí:
usted proporciona cuidados constantes y continuos a un miembro de la familia o del hogar que tiene COVID-19
Cuando contestar que no:
usted no proporciona cuidados constantes y continuos a un miembro de la familia o del hogar que tiene COVID-19
12. Did you become the breadwinner or major support for a household because the head of household died as a direct result of COVID-19?
Cuando contestar que sí:
se convirtió en la principal fuente de ingresos de su hogar y no puede encontrar trabajo
Cuando contestar que no:
se convirtió en la principal fuente de ingresos y está empleado durante esta semana
no se ha convertido en la principal fuente de ingresos tras el fallecimiento del cabeza de familia
13. Were you offered the option to telework (work remotely or from home) for your usual number of hours with pay during the COVID-19 public health emergency?
Cuando contestar que sí:
ha recibido la opción de teletrabajar y puede hacerlo esta semana durante su número habitual de horas con sueldo
Cuando contestar que no:
no recibió la opción de teletrabajar por parte de su empleador habitual
14. Can you telework (work remotely or from home) this week for your usual number of hours with pay during the COVID-19 public health emergency?
Cuando contestar que sí:
puede teletrabajar su número habitual de horas con sueldo esta semana
Cuando contestar que no:
no puede teletrabajar su número habitual de horas con sueldo durante esta semana
no puede teletrabajar porque está cuidando a sus hijos de forma constante y continua mientras las escuelas o guarderías están cerradas
no puede teletrabajar debido a la naturaleza de su trabajo (por ejemplo, si trabaja en la construcción o en un restaurante)
no puede teletrabajar debido a violencia doméstica, violencia sexual o el acecho
15. Did you apply for or receive any private income protection insurance or supplemental unemployment benefits?
Cuando contestar que sí:
ha recibido un seguro privado de protección de ingresos de proveedores como Aflac, MetLife, The Standard o Guardian
ha recibido prestaciones de desempleo adicionales, en las que el empleador suele pagar las primas cuando el desempleo se debe a una enfermedad, una lesión o una reducción de la fuerza laboral
Cuando contestar que no:
no se aplica ninguno de los puntos anteriores
16. Have you applied for or received unemployment benefits from another state or the U.S. Railroad Retirement Board in the last 12 months?
Cuando contestar que sí:
ha recibido prestaciones de desempleo, incluyendo extensiones de cualquier estado o de la Junta de Retiro Ferroviario
Cuando contestar que no:
no ha recibido prestaciones por desempleo, incluyendo las extensiones de cualquier estado o de la Junta de Retiro Ferroviario
17. ¿Se le ha denegado la continuidad de los beneficios por desempleo porque se negó a regresar al trabajo o aceptó una oferta de trabajo en un lugar de trabajo que no cumple con las normas de salud y seguridad locales, estatales o nacionales directamente relacionadas con el COVID-19? Estas incluyen, entre otras, las relacionadas con el uso de cubrebocas o caretas, las medidas de distanciamiento físico o la provisión de equipo de protección personal de acuerdo con las normas de salud pública.
Cuando contestar que sí:
sus prestaciones de desempleo se interrumpieron porque se negó a volver a trabajar en un centro de trabajo que no cumplía las normas
sus prestaciones de desempleo se interrumpieron porque no aceptó una oferta de trabajo en un centro de trabajo que no cumplía las normas
Cuando contestar que no:
ha seguido recibiendo prestaciones de desempleo
sus prestaciones de desempleo se interrumpieron por motivos distintos a los anteriores
18. ¿Proporcionó servicios a una institución educativa o agencia de servicios educativos y está desempleado o parcialmente desempleado debido a la volatilidad en el horario de trabajo causada directamente por la emergencia de salud pública del COVID-19? Esto incluye, entre otros, cambios en los horarios y cierres parciales.
Cuando contestar que sí:
usted ha prestado servicios a lo anterior, incluidas las escuelas K-12 y los colegios privados y públicos; Y ADEMÁS
usted está desempleado o parcialmente desempleado debido a cambios en su horario de trabajo como resultado directo de la emergencia de salud pública COVID-19 (el estado considera si usted está reclamando una semana que está entre o dentro de los ciclos académicos y, si lo está, si tiene un contrato o un trabajo que está disponible para que usted regrese en el siguiente año o ciclo académico).
Cuando contestar que no:
ha prestado servicios a los mencionados anteriormente pero aún está trabajando o tiene un empleo parcial
no prestó servicios a ninguno de los empleadores mencionados
19. ¿Es usted un trabajador y le han reducido sus horas o lo despidieron como resultado directo de la emergencia de salud pública del COVID-19?
Cuando contestar que sí:
sus horas se redujeron debido a la emergencia de salud pública COVID-19
ha sido despedido como resultado directo de la emergencia de salud pública COVID-19
ha sido despedido pero no ha podido solicitar las prestaciones regulares de desempleo porque no ganaba lo suficiente para tener derecho a ellas, porque había sido descalificado anteriormente o porque ha agotado todas las demás prestaciones que tenía a su disposición.
Cuando contestar que no:
sus horas no se redujeron debido a la emergencia de salud pública COVID-19
no ha sido despedido
Created for states using the FAST UI platform and tested for usability, these prototypes can be used as a blueprint for any state, regardless of platform or vendor.
Display hover states when user mouses over buttons
Yes/No (Si/No) Buttons should be above the fold (640 px)
Made helper text grey to de-emphasize the hierarchy
Made the Yes/No (Si/No) buttons black and bold to emphasize these buttons
If user selects Yes:
Display “From" and “To" date picker fields
Display button to add more dates upon entering date in “To” field
“Next” button disabled until user fills in all dates in date picker
If user selects No:
“Next” button becomes active
Outreach content that can be used for both digital and paper notifications to help claimants understand why they are being contacted and designed to mitigate for the perception of fraud/spam.
Subject
Aviso: usted puede ser elegible para re aplicar a la asistencia de desempleo debido a la pandemia
Body
A principios de este año, el Departamento de Trabajo de EE.UU. amplió las disposiciones de elegibilidad a los beneficios federales por desempleo debido a la pandemia para incluir tres nuevas razones relacionadas con la COVID-19 bajo las cuales un individuo puede autocertificarse:
Individuos que se niegan a volver a un trabajo que es inseguro o a aceptar una nueva oferta de trabajo que es inseguro;
Ciertos individuos que prestan servicios a instituciones educativas o agencias de servicios educativos;
Personas que hayan tenido una reducción de horas de empleo o un despido temporal o definitivo.
Estas disposiciones se dan de forma retroactiva al inicio del programa de los beneficios federales por desempleo debido a la pandemia, aunque las personas que no presentaron una solicitud inicial el 27 de diciembre de 2020 o antes, se limitan a las semanas de desempleo que comienzan el 6 de diciembre de 2020 o después.
En base a estas nuevas disposiciones de elegibilidad, ahora podría ser elegible para los pagos de los beneficios federales por desempleo debido a la pandemia. El Departamento de Seguridad de Empleo pronto le notificará por correo electrónico cuándo podrá autocertificarse en el sitio web.
Lo que esto significa para usted
Usted debe certificar en línea en esd.wa.gov para tener la oportunidad de establecer si es elegible cuando nos volvamos a comunicar con usted. Al no hacerlo, podría no enterarse si anteriormente se le negó incorrectamente los beneficios federales por desempleo debido a la pandemia y si se le deben semanas de fondos.
Gracias y esperamos tener una respuesta pronto.
Subject
Recordatorio: Fecha límite para la autocertificación de los beneficios federales por desempleo debido a la pandemia
Body
Este correo electrónico es para recordarle que se acerca la fecha límite para establecer la elegibilidad de los beneficios federales por desempleo debido a la pandemia.
Por favor, ingrese a esd.wa.gov para responder con su información antes del [fecha].
No espere. Esta es su oportunidad para determinar si anteriormente se le negó incorrectamente los beneficios federales por desempleo debido a la pandemia y si se le deben semanas de los fondos de los beneficios federales por desempleo debido a la pandemia.
Subject
Último recordatorio: Fecha límite para la autocertificación de los beneficios federales por desempleo debido a la pandemia
Body
Esta es su última oportunidad para establecer la elegibilidad de los beneficios federales por desempleo debido a la pandemia en nuestro sistema antes de la fecha límite de [fecha].
Ingrese ahora a esd.wa.gov para responder con su información antes del [fecha].
Una vez que pase su fecha límite, ya no podrá determinar si anteriormente se le negó incorrectamente los beneficios federales por desempleo debido a la pandemia y si se le deben semanas de los fondos de los beneficios federales por desempleo debido a la pandemia.
By Alyssa Levitz
Unemployment Insurance agencies have been experiencing an unprecedented and overwhelming volume of telephone calls related to unemployment insurance during the COVID-19 crisis. There are simply not enough phone operators to handle the volume of incoming calls, and callers are having poor experiences (e.g., dropped calls, long wait times).
Having a complete "suite" of customer support tools not only helps agencies control costs associated with some of the higher-touch support channels, but it also helps their constituents by meeting them where they are. Phones (and when safely possible given COVID restrictions, in-person appointments) are incredibly important for a speedier resolution to complex cases -- and for people with limited sight or reading capabilities. They also provide a mechanism for those with limited or no access to the internet to get their questions answered and cases resolved.
Keeping phone lines and in-person appointments in place are important for reducing the impact of the digital divide. But without a suite of customer support tools, everyone relies on the phone and ends up crowding out the people who need the phone to do anything at all with their claims or questions.
We have done deep dives into some of the solutions that can be complementary to existing customer support systems:
An unknown number of legitimate claimants are being prevented from receiving the benefits for which they are eligible. This marks a failure of the UI system that has not received as much attention as it deserves.
There is currently no way to tell the difference between someone who gave up because they were a criminal and knew they couldn’t pass the step, and someone who gave up because they didn’t understand or have the required documentation to move on.
But we can--and should--tell a different story: one that gives as much weight to how easy these systems are for the rightful claimants to pass through as it does to how good they are at catching criminal actors.
It is currently in the interest of both identity proofing vendors and the UI agencies that contract them to conflate unstarted and incomplete applications with fraud. The vendors’ performance looks more impressive if they don’t counter the perception that everyone who doesn’t get positively identified is fraudulent. UI agencies, under intense scrutiny from the public and US DOL, are able to look more effective.
The agencies should work with the vendor(s) that provide their identity verification solutions to ensure that the unstarted and incomplete applications are tracked and reported separately from those whose incomplete or completed applications had a positive indicator of fraud detected. That data from the vendor should be made available on a user-by-user basis to the UI agency so they can connect it to their own claimant data and break it down by race.
Distinguishing this “false” rejection rate from the overall rejection rate of these systems is a key step toward improving benefits participation rate.
While the document+biometric verification products will have gone through usability testing within the vendor's company, UI agencies should usability test the end-to-end experience themselves or in partnership with the vendor.
Whether the identity proofing is included in the initial application, triggered by a change on the account, or requested as part of a backlog-clearing process, it exists in a context beyond what the vendor could've usability tested independently: e.g., surrounding website interface, digital notification or physical letter asking them to complete the identity proofing process. Here are some key questions to ask about the experience:
How is someone notified that they need to go through additional identity verification?
Does it make clear what people should do if they don’t have a photo ID or have difficulty going through the vendor’s process?
What methods are used, and are those effective for the population in question? (Have you considered using SMS?)
Has the notification content been usability tested in all languages? (Do people understand what is being asked of them, and the importance of them following through?)
Is the claimant reminded to start the process if they don’t ever begin? (They should be.)
What languages is the vendor’s user experience available in?
If someone starts but does not finish the application:
Is there a clear way from the vendor’s user experience for someone to get direct agency help if they are unable to provide what the process requires?
Is the claimant reminded to complete the process if they begin but do not finish? (They should be.)
By conducting their own usability testing, UI agencies have the opportunity to make sure that the experience has been tested by a representative sample of people based on the demographics of the population seeking UI benefits. If anything is found to be unclear, the agencies should work either on their own or with the vendor to ensure that the experience or communication is improved.
If these patterns are at all present in the 1:1 matching use case between a selfie (which isn't necessarily going to be high quality or well-lit) and an ID photo, it would contribute to racial disparities in benefits participation rates.
as Google Documents and download in your preferred format.
as Google Documents and download in your preferred format.
But, again: The fact that someone was stopped from receiving benefits does not mean that they were rightly stopped. Agencies and vendors should be careful with their language and not .
How long does someone have to complete the identity proofing process? Does the data support this time period? (California increased the period .)
study co-authored by researchers at the MIT Media Lab and Microsoft Research, lighter-skinned men were more than 40x less likely to be misclassified than darker-skinned women in common face recognition software. A to that paper said that they made significant improvements to their system, but lighter-skinned men were still 13x less likely to be misclassified than darker-skinned women.
The found that when image quality is low and you're trying to match a photo to one of many possibilities (1:N), "false negatives are generally higher in people born in Africa and the Caribbean, the effect being stronger in older individuals."
Ultimately, more information is needed on how false rejection rate differs by skin type. ,* a company with a document+biometrific verification product, has been transparent about how its , but there doesn't appear to be data on false rejection rate by skin type from any of the major players in the space.
UI agencies should ask the vendor they are using for document+biometric verification for these performance metrics, and if there is any discrepancy between performance for different skin types, they should ask how the vendor is planning to improve. ,* another company that provides this product, published "," which is a succinct and useful list to reference.
* Note: USDR is neither affiliated with nor endorses any vendors. Additionally, at the time of this publication, neither onfido nor Jumio has been evaluated as part of USDR's , but that shouldn't be seen as an indicator of their suitability for UI agencies' needs.
Here’s how the state agency describes their high-level needs, and USDR’s notes about that requirement after looking at the first dozen vendors or so. To see the evaluation matrix along these and other dimensions, download this spreadsheet:
SaaS-based scheduling tool.
Ability to define # of appointments available based on # of agents.
As we look at the different tools, it doesn’t appear on the surface that any offer this particular functionality. However, 1) calls with sales reps could verify this, or 2) we can think about this a different way. E.g., you may have to figure out how many appointments you can manage per slot, and the question becomes: how easy is it for you to set and update that “schedule”?
Ability for customers to self-schedule their own appointments.
Almost all the evaluated providers have the ability for you to “embed” the tool within your own webpage; some additionally have the option for you to create your own experience and use their API, or for you to use a landing page that they provide for you.
Ability to prevent duplicate appointments being set by same individuals.
No tool mentioned this as a feature, but we can ask about it in sales calls.
Note: if the system does a good enough job sending a confirmation email/SMS right after the person books the appointment, we might not see as many duplicate bookings.
Alternative: We might also be able to think through a solution to put “on top” of the vendor’s implementation. The following is an example of a solution that could work depending on the UI system’s technical limitations:
Scheduling appointments would be available only after logging in. If the tool has webhook/API support, whenever someone schedules an appointment, it would let the state's UI system know that it had happened, and then remove the option for that person to schedule another appointment, showing instead “Appointment already scheduled” or something to that effect.
You could take this a step further and have the tool send you back information about when the person’s appointment actually is to display that and give them a clear way to cancel or reschedule.
Confirmation emails/reminders to customers:
Almost all tools have both confirmations and reminders. The biggest question is email vs. SMS. SMS costs more from many of the vendors but may be more effective at reducing no-shows.
Ability to customize forms for collecting information
The tools labeled as “candidates” in the spreadsheet all have this ability; however, some only have this ability in different pricing tiers. We’d want to clarify what kind of intake data you want and run that past the sales folks to understand the feasibility of your specific need.
Ability to see all callbacks scheduled and make agent notes and set indicators for those we were unable to reach:
All the vendors marked as “candidates” have some level of reporting, though it can vary by pricing tier.
It was not apparent with most of the vendors how you might annotate individuals as un-reached; that might be something that you integrate with your CRM or have an agent manually look up and note in another tool at their disposal. Vcita has this capability in its higher-cost tiers, as does Nemo-Q (both vendors used by state DMVs).
Form localization support, including callbacks in the chosen language:
I need to go back and look for localization support of the intake form and calendar interface.
Re: callbacks in the chosen language: this could be figured out by collecting that information as data in the intake form.
Ability set workgroups so callbacks can be schedule by certain teams:
This gets to the ideas of “users” and “access”, which are ill-defined in many descriptions of the tools’ features. This will likely become clearer when we’re able to talk to salespeople and describe our scenario to get feedback.
Various reporting capabilities for metrics/quality control:
Again, we’d want to know more about what specifically the agency is looking for to evaluate this furhter.
Additional desirable features:
Some form of integration capabilities with CRM:
Many of the candidates do have this via Zapier, which is another app that helps provide the connection between different systems.
Ability to standup multiple instances to take advantage of different divisions’ needs:
Most of the tools offer this in some manner, though some might be sleeker than others.
Dashboard view of callback status for staff performance/monitoring:
This seems beyond the scope of many of the tools, but it is something that we could ask about in a sales call. Likely the first sticking point would be the fact that people wanting a call back aren’t scheduling an appointment with a specific individual.
Evaluated:
Acuity Scheduling
Appointy
Calendly
DocPace
FlexBooker
Juvonno
LumaHealth
Microsoft Bookings
MS Dynamics
QLess
Schedulicity
Setmore
Solv
Square Appointments
Vagaro
vcita
You Can Book Me
Not evaluated:
Appointlet
AppointmentPlus
EZnet Scheduler
Genbook
Marketing360
Qmatic
SimplyBook.me
TimeTap
This report combines research from many different sources, which are summarized below. It was compiled by Alyssa Levitz, Waldo Jaquith, and USDR but the underlying work was also done with the help of multiple volunteers and civil servants who can’t be publicly listed, but deserve credit for their work all the same.
We are indebted to the research that has come before us and are deeply grateful to those who have spent time speaking openly with us and given us the opportunity to support them.
USDR partnerships on 16 projects with 10 states, including:
A Discovery Sprint to help a state understand their backlog
Vendor analysis for online appointment scheduling
Interviews with:
Subject matter experts
NASWA
DOL and UI Leadership in 6 states, referred to as States A - F in this report
California EDD and Nevada DETR Strike Team Reports on the UI systems
Others’ detailed research:
“Centering Workers--How to Modernize Unemployment Insurance Technology” by Julia Simon-Mishel and Maurice Emsellem from Philadelphia Legal Assistance; Michele Evermore and Ellen LeClere from the National Employment Law Project; and Andy Stettner and Martha Coven from The Century Foundation.
“Story-driven experience research on pandemic unemployment” by Dana Chisnell of Project Redesign at the National Conference on Citizenship and Tara McGuinness of New America’s New Practice Lab
“Unpacking Inequities in Unemployment Insurance” by Monée Fields-White, Vivian Graubard, Alberto Rodríguez, Nikki Zeichner, and Cassandra Robertson from New America’s New Practice Lab
Current UI systems are not designed to simplify the claimant’s experience. Application questions are complex and difficult to understand. People are nervous and worry that clicking the wrong option will result in losing their benefits by mistake. They don’t always provide the correct information; claimants might send over too much information, which can confound and complicate decision making, or not enough information, which lengthens the process. Online applications are not always mobile-optimized and frequently aren’t translated into the languages needed for the state’s population.
Agencies should explain the outcome of actions that the claimant takes or of answers they provide.
When additional information is needed, agencies should be very clear about what that information is and why it is needed. Uploading attachments should be available to claimants in the context of a targeted request for information from the agency, or as part of an issue creation/appeals workflow that the claimant initiates.
Create a substantive, accessible claimant portal.
Go for a professional look.
Make your website mobile-optimized.
Design a sensible password reset process.
Make online and mobile systems available 24/7.
Automatically save incomplete applications, and provide a warning before timing out.
Allow customers to choose email or texting as a communications method.
Permit customers to email in or upload documents from a computer or mobile device.
Avoid automated decision-making.
Use plain language and smart questioning.
Translate the application and other online materials into Spanish and other commonly spoken languages.
Minimize the paperwork burdens associated with work search.
Some systems were not built for increased web traffic load. This has resulted in some states trying to limit what days or times individuals can apply or do their weekly certification in order to balance web traffic load.
Agencies should pt in place scaling plans and determine what will change in the system when there are large spikes in traffic. Take into account DDOS attacks as well as bot traffic.
Not everyone has access or sufficient digital skills to use the website.
"The people impacted by this pandemic are blue collar, they’re in socioeconomic groups that may be put at disadvantage as particular industries, educational background, they’re at a disadvantage due to complexity of system and ability to apply for claims, plus a language barrier that’s highlighted." -- State B
Agencies should create a process for phone applications and certifications and make sure that a phone line is available specifically for that purpose. This phone line should be separate from getting help with online applications or handling existing application or status issues.
Agencies should maintain a process for handling paper applications.
When possible to do safely, agencies should consider in-person “pop up” offices in targeted locations can help provide assistance. Think through the time commitment and resource constraints before setting expectations with the public.
Agency websites should eable claimants to schedule phone callback appointments to preemptively avoid long hold times.
Unlike most of the online interactions people have, UI systems frequently aren’t able to tell you right away what the status of your application is. If an issue is detected, or the application needs to be processed by the mainframe in a batch overnight, a claimant is left wondering what’s going on. When the system doesn’t tell them what is going to happen next or on what time frame, then claimants have undue anxiety and confusion, as well as calls to the help center.
“You’ll enter a claim but it doesn’t update until tomorrow. We have batch processes with the mainframe, so it’s not real time, it’s as of 24h ago. The dashboard might still say that there’s a problem even if the issue is resolved.” --State B
"Unemployment works differently from every other experience you have in the world, where you are approved until we change our minds about it. That can happen at any time. People assume they’re walking into a system where they’re going to fill out paperwork and payments are going to start. That’s just not an accurate representation of how unemployment meaningfully works. It’s an ongoing relationship where we’re constantly checking in. The term “eligible” here means something different to our customers than it does for us. It’s something we’re checking every week before issuing a payment" -- State F
Where possible, agency sites should seamlessly transition users in between processes. (E.g., if someone’s application is flagged immediately upon submission because more information is needed about their self-employment, go directly from that submission to a form/questionnaire.)
When information is submitted, agencies should let claimants know what will happen next and when they should expect to hear back.
Agencies should use in the tools to help claimants better understand what is expected of them. US DOL should provide plain language guidance for common scenarios that States can use.
From “”, agencies should update their applications according to the following:
State UI systems often have no digital status-checking method as it can be hard for the website to display succinct information across multiple backend systems. This requires claimants to call customer support to find out what is going on with their claim.
"People also think of many of these apps similar to a bank loan. They expect to put in information and then see a status. Our system and laws and regulations, it’s not that simple. So there’s a disconnect between what people expect to see and expect to be able to understand about their app and their benefits, and what we’re able to give them based on the restrictions in our system" - State B
“People at the call center are looking at 7 different systems: this one says you got paid this way, there’s [an issue over here, so it’s] hard to give clear claim status. We want to give people a weekly or monthly statement of what you got paid. Doing that right now combines a bunch of system and not easy to do.” -- State C
"It’s not obvious whether you’re waiting for the state to do something, or if the state is waiting for you to do something. You don’t have visibility into what’s going to happen next. Anybody with much anxiety is going to call in order to confirm information" -- State F
Agencies should iterate toward a complete status-checking experience - they shouldn’t try to do it all at once. E.g., it’s impactful to show someone their initial application status even if you can’t yet show weekly certification status or vice versa, or to show that a weekly certification was deemed eligible even if you can’t yet show how much/whether they have been paid.
Agencies should consider providing claimants with a way to check their status via automated phone, web chat, or SMS in addition to (or even as a step before) being able to check their status on the website.
The language used to convey application status is often meaningless to claimants; “application pending” could mean any number of different things
"When we send out letters to use this language like adjudication or pending issues or ICON, no one knows what that means. One of the problems is there are so many different details that we need to work out to process their claim to describe them in easily understandable language." -- State A
"There are generic statuses, like ‘pending.’ But you wouldn’t know why. If we are waiting for missing wages, the claimant would have no idea what’s happening. This is a big pain point for the claimant and for us. It generates a lot of calls. The Domino's app approach (you can see the pizza being made, you can see it being delivered) would be a lot better. Right now it’s a black hole." -- State F
Agencies should communicate in plain language what the status is; statuses should be descriptive yet concise. US DOL should provide guidance that all agencies should use.
Agencies should ensure that their claim and payment statuses are contextualized within the whole process, e.g. a progress bar. Statuses should also include the following supporting information:
Any decisions that have been made, and ideally the ability to begin an appeals process on that decision from that point.
What issue(s) are set on the claim, as well as what (if anything) is needed from the claimant; claimants shouldn’t be left wondering if they’ve missed a communication.
What payments have been disbursed, if any.
Updated March 14, 2021
Many states limit how funds can be received, including only providing benefits via a prepaid debit card. Prepaid debit cards are targets for both physical and digital fraud, and enable banks to withhold or retract funds that they think were obtained fraudulently (without being asked by the state to do so).
"There’s also a challenge with new filers in the system who may have had a paid family leave or disability claim in the past four years. Those UI payments would go onto that same card. So if they lost that card or threw it away, there can be some challenges where the claims are crossing over to same benefit payment method but claimant doesn’t realize that and so there’s challenge" -- State B
People ask for Venmo more than they ask for checks --State B
"Direct deposit is great for people who are housing-unstable. Otherwise getting the debit card to people is a challenge. There are some non-traditional banks that act as direct deposit, and that’s helped. You can go into Walgreen’s and get a card that acts as a direct-deposit mechanism, so there are options for the unbanked." -- State F
Arizona, California, Maryland, Nevada, and Oklahoma should enable direct deposits of UI benefits without having to mail in a voided check or first get a debit card.
Claimants UI system accounts can be hacked and their bank account info changed so that someone else gets their benefits.
"We ask them to validate their identity using the exact same process every time we have an identity question. The process earlier, they’d do the same thing later, you can imagine how frustrating in the first place, and you had to do it again for opaque reasons, it might seem like we’re putting you on, in the nicest version of that. It’s a particularly frustrating point of the system for me. " -- State F
"The changing of bank accounts has been a strong indication of possible fraud, so there’s a two-day hold if people change bank account information. This can result in people in need of money being delayed." -- State F
Agencies should assume that bank account changes are hostile actions and require authentication beyond simply logging into the account. Be clear with the claimant about what is happening at that time and how to confirm legitimate changes.
In progress
This is a challenge for validating monetary eligibility as well as “situational” eligibility. Whenever someone files a UI claim, the employer who will be charged for the claim must be informed.
Many employers get all information from the agency and can only respond by paper and mail (this is inherently slower than digital processes, and has been exacerbated in the last year by mail delays)
Even for those employers that use more digital tools, they may not get notified of a new claim by a former employee, and so everyone must rely on that person proactively going to the site frequently.
If employers don’t want to challenge the claim, they aren’t required to respond. If they don’t respond within a certain period of time (depending on the state), then the claim is determined to be unchallenged and can proceed. However: several states, including PA and MA, do not have an automated system to move claims to the next step if employers don’t respond. Thus, in times of high volume, claims can get stuck in this step for a very long time.
US DOL should launch a specific initiative with states and NASWA to co-develop solutions to improve and digitize agency interactions with employers
When employers see notification of a claim, there’s a form they use to respond back to the agency. In many states, this was not updated to be able to mark identity theft, which decreased efficiency and/or made it harder for the UI agency to tell what was going on with a particular claim.
"Enhanced wage records - we’re working with the chamber foundation and other groups to come up with standardized wage records not just for unemployment but also many other DOL interests as well. In this pandemic, it’s amazing how many times I hear that their employer submitted information incorrectly. Everything in this situation is bigger than usual, but this is extremely frustrating for us, for me personally, the claimant didn’t do something wrong, it was their employer who has presuming they’ve been doing things right all these years but actually been incorrect." --State B
The experience rating system as it stands today influences (and incentivizes) employers to contest a former employee’s claim to unemployment benefits. When an employer challenges a claim, fact-finding is initiated. In this fact-finding, the UI agency must get information from the employer and from the would-be claimant; an expert in state UI law is required to conduct the fact-finding to be able to make a determination. Thus, there is a cost to UI agencies associated with every employer challenge.
This process can be lengthy, with both claimant and employer needing time to be able to respond, sometimes with many documents that the agency expert must wade through. In the case where the claimant is found to be eligible for benefits, this process has simply served to increase the amount of time it took to first payment.
Additionally, employers are at a strategic advantage when it comes to these challenges, in a way that interferes with the principle of ensuring that all eligible claimants should be able to receive benefits. Employers frequently have lawyers or contractors hired specifically for the purpose of making these challenges (Equifax and ADP are major players in this space). In contrast, would-be claimants frequently do not have the resources or knowledge to state their case effectively in contrast. Claimants are unlikely to have the funds to hire a lawyer, and may not know that legal aid organizations could help them (and even if they did know this, the organization might not be able to take their case in a timely manner).
US DOL should recommend to States that their experience rating system incorporate the percent of claims against an employer that are challenged, and of those that are challenged, what percentage of them are decided in the company’s favor, compared against some baseline. That baseline should account for the fact that right now, it is likely that too many claims are being “successfully” challenged (though of course there is always going to be some number of claims found in the employer’s favor). This normalized index of each employer’s experience with the UI appeals process can be incorporated into the scheduling of tax rates for one employer in accordance with the relative experience of other employers.
We have "daily ‘Severity Level 1’ incidents in IT, because our infrastructure has a hard time keeping up." --State C
“The challenge is, the law passes, the president doesn’t sign it, then he signs it, but what people hear is that there’ll be $300 more in your account tomorrow. And that’s not how it works, we have to reprogram everything." --State D
“Normally in a business you’d have two versions of this program so you could make updates to the program on the one while being live on the other - we can’t switch over like that. We can only take pieces offline, we’ve been trying to do a security update, and we try to put it in, and something else goes wrong, we have to back it out, take our whole system offline for half of a saturday every week.” --State D
“Right now is the best the mainframe is ever going to be"" 5 minutes from now it may be worse. [...] My claims agents work from a green screen, like Oregon Trail. It’s Oregon trail every single day." --State E
“We tried to move quickly and the changes aren’t embedded or tested as they should be in normal times cause you need to get them out.” --State F
US DOL should provide guidance on what it means to modernize the system based on solving agency and constituent problems, not just on the need to get off mainframes and migrating data. Funding for modernization should be incentivized to get states to use modular contracting and agile software development.
US DOL should work with the US Legislature to find more consistent funding sources than large, one-off grants from stimulus packages.
US DOL should establish a USDS team reporting to senior leadership that exists independently of the areas responsible for audit and enforcement to work with states to build a “no-blame” environment of trust and address urgent pandemic issues - including conducting discovery, enabling sharing of concerns, and creating a forum to get feedback on potential actions by the federal government.
Use deployed cross-functional USDS teams that establish good working relationship with states to iteratively transition from ‘rapid response’ to ‘modernization’ at the right time/pace so that shared services can be built and deployed together with states
Offer to conduct rapid-review sprint to provide recommendations to the state and inform the development of a best practice guide
Open offer to deploy cross-functional USDS teams to states to implement items from the recommendation sprint
Offer to conduct user interviews with state claimants to provide insights to improve the delivery of UI benefits in the state
Offer to provide states with procurement/contract support, including the review of existing contracts and issuance of new contracts
“If they pick up wages in another state, there is a hold put on the claimants; Send request to ICON & ICON sends request to other state; Really depends on another state; Done by hand, no automation; Takes forever. Because each state has a different process, it causes more delays. Can this process be standardized to flow faster and easier?” --State A
US DOL should launch specific initiatives to co-develop solutions with states to share interstate wage data. Establish a working group that includes states and NASWA to address this. ICON (the current system) digitizes requests between states but does not help with tracking or ensuring a timely response. US DOL should further partner with select states to iteratively build and test shared service components and/or reference implementations that will inform country-wide guidelines for UI system improvements
US DOL should update guidance to enable people to receive benefits from home state first, and make adjustments when the data from other states comes in. Learn more
"[I wish we had a] Joint case management system [across benefits programs]. [State] is very fragmented for someone to get services. We want to use tech more efficiently and effectively." --State A
Our state has “issues knowing if people are receiving other benefits from the state that might make someone ineligible.” --State B
State agency and/or US DOL: If any data needs to be checked with another agency, then make sure that the other entity is prepared for the incoming requests. Ideally, you will be able to establish an expected response time so you could build your surrounding processes accordingly. Digitizing the request to the other agency is important to making sure it doesn’t get lost, but if the other agency’s underlying process is manual, then the digitization won’t have that much impact on timeliness of claimant’s benefit receipt.
Agencies that need to check someone’s military veteran status should build their process around the Veteran Confirmation API, rather than using separate manual processes
Agencies: For any group of people that may need additional or special verification, make sure there is an accessible, mobile-friendly way to upload documents (including image files) to avoid requiring manual/paper process for the population"
US DOL should work with OPM to build an API to verify that someone was a civilian employee and what their last pay grade was
State UI agencies are not set up well to validate wages earned through self-employment because those are definitionally not reported by an employer, which is the standard source of verification for UI systems. To the degree that PUA may persist beyond the pandemic, it’s important to think about improvements to this wage verification system.
"The IRS should be giving us everything, it’s a federal program, this shouldn’t have been with DOL. We don’t have this information, it was shoehorned into us to the detriment of everyone else." --State B
“PUA was a perfect storm of good intentions and bad policy. [...] Are you trying to break our system?” --State C
- “When DOL said you guys aren’t giving the right monetary determinations for these people, you need to start doing this. OK great, can you help us figure out how to do this, what would be your advice on how to do this quickly? We want to be in compliance, but it’s human beings doing math.” --State E
Similar in concept to the recommendation for Challenge 2, state UI and US DOL should work with state and federal treasury and IRS to get this data directly and/or leverage the information already gathered for other benefit programs. While APIs/agreements for working with the IRS almost certainly don’t (yet) exist, states with income taxes should work with their treasury departments to come up with an efficient process to validate wages
US DOL and the executive branch should make and communicate a decision about the future of PUA beyond the pandemic to inform how much state agencies should invest
Unemployment insurance agencies operate under a “threat model” for improper payments of detecting claimants who lie or otherwise make mistakes on their applications and weekly claims. When substantial federal money began flowing into these systems with the CARES Act, the agencies quickly became targets of criminal digital fraud syndicates. These criminal rings have primarily been using stolen identities to claim others’ legitimate benefits or, in the case of Pandemic Unemployment Assistance, to create fully illegitimate claims.
"There have been more data breaches over the last couple of years where accurate information is highly available to folks. In the past, if you have all the right info for someone, our system wasn’t built for stopping you. In most cases, fraud meant that someone was getting unemployment incorrectly - accurate PII out there has made everything more difficult." -- State B
"The problem we had previously was that someone who stole someone’s ID has all their information, they would be able to get past the software, so what we found is that our ID verification was only stopping people who were the right person who couldn’t answer the questions right. We were stopping the wrong people." -- State B
"Right now, there’s been a pretty substantial correction toward making sure that criminal rings can’t come in and file claims on behalf of people." -- State F
Agencies should ensure their threat model is up-to-date: what kind of fraud or theft will your system be susceptible to? (Identity theft to steal others’ benefits, or “benefit theft” to try to get more than you are otherwise eligible for) Agency identity proofing systems shouldn’t rely on Knowledge Based Verification. Look to have a NIST Identity Assurance Level 2 system. To learn more, see USDR’s Identity Proofing for UI Agencies report
In UI systems, there is a balance between finding fraud and getting benefits out in a timely manner. Right now, a lot of fraud is slipping through and benefits are severely delayed. As we work to shift this balance, we need to ensure that we reduce -- not broaden -- racial inequities in UI.
“By focusing on overpayments, the [US DOL] is not held accountable for determining other errors created by the state, or even an employer, when it comes to UI payments—such as underpayments—which are also payment inaccuracies. The [accuracy] map also doesn’t reflect cases in which eligible UI applicants are erroneously denied benefits.” --Unpacking Inequities in Unemployment Insurance
“ID verification, the tools we use have the biggest weight against, have the heaviest burden on, low income people. If the tool is a driver’s license, how much more difficult that becomes for low income people." --State B
"[We have] no way to break down fraud vs. not having a driver’s license. We’ve seen an uptick in phone filing because of this.” -- State B
"These systems, the larger populations get prioritized because you get the biggest result but it’s not the right way to think about things. We are trying to shift and think about smaller populations but more difficult time getting served and get them served with priority "" -- State F
With the support from US DOL, agencies should find and mitigate inequitable impact of identity fraud detection flags like IP address, multiple uses of the same physical address, and restrictions on characters used in people’s names. Learn more
With the support from US DOL, agencies should increase accountability for ensuring rightful claimants make it through the system, like usability testing the entire identity proofing process. The US DOL can set standards and best practices for use. Learn more
US DOL should reevaluate KPIs and their relative importance to ensure timeliness and equity are valued and improved upon.
US DOL should re-release Replacement Rate data in easily accessible format
"With the claimant, because we’ve had so many issues with stolen identities used to file fraudulent claims, when the real claimant goes to file, we have to undo everything that has happened - new username against SSN, if payments got issued, we’ve identified a number of claims beforehand, but some need to be cleaned up so that the payments aren’t charged to the claimant or employer. If the fraudulent claim was issued a debit card, the bank is going to have to cancel that, it’s tied to the SSN. It’s not an easy process, especially with the volumes we’ve been having" --State B
Agencies: when the threat model indicates high likelihood of identity theft as a means to obtain benefits, plan ahead for how the rightful claimant can come in after the criminal and actually receive their benefits without significant delay (even if the stolen benefits haven’t been recouped from the criminal).
Additionally, the states acting independently leaves the slower-acting states vulnerable: As states tighten their fraud detection efforts, criminals move to other parts of the system (State A vs. State B, standard UI rather than PUA, or hacking into accounts to change bank info rather than submitting new applications)
“This has been a failure of imagination at every turn. People have not appreciated the scale and scope of this. We’re getting crucified in the media because of things that are beyond our control. To say this is demoralizing is an understatement.” -- State C
We would be "better served if national requirement and participation in the common data sharing were required" -- State F
"Federal prison data that SSA has and NASWA has been trying to get it and build it into integrity center and they can’t get the data sharing agreement figured out with DOL. We should just solve these so individual states aren’t having to do independent efforts that will contribute to fraud nationally" -- State F
Legislature: The Social Security administration should be given resources and a mandate to make its online, live, SSN validation service available to state agencies both directly and through Login.gov.
US DOL: should set up a “fusion center” to set up and facilitate an ongoing information exchange, so States can learn from each other about what fraud tactics are being used and how to combat them. As part of the project, define the process (and technologies) that people will use to report identity theft, and that the agency will use to act on that information (including law enforcement involvement)
US DOL in partnership with other federal agencies: US DOL should secure the use of the CMS Data Hub to provide a shared eligibility determination service, and as an additional data source for identity validation. This makes the process much more efficient for many who won’t have to manually re-enter information for every new program; however, you’ll have to at the same time ensure that those who aren’t in the partner program still have a pathway to apply. The Missouri Benefits Enrollment Transformation report has many transferrable recommendations, pp 28-29 in particular.
Federal GSA in partnership with US DOL: DOL should pay for Login.gov use for any states that wish to use it, and provide incentives for states to do so
State agencies and US DOL: Programs administered by other agencies may have already done the identity validation you need, or could use the ID proofing that you do. Existing auth systems (e.g., with SNAP or Medicaid)) should be leveraged.
Visual accessibility: There are many best practices when it comes to accessibility for the visually impaired. The chat tool should have high-contrast and work with screen readers at the very least. You should not choose vendors that don't meet these requirements -- or you should push your vendor to update their tool for ALL customers to begin meeting these requirements.
Translations: Based on the demographics of your state, there are likely one or two other languages that your site -- and all corresponding tooling like chatbots -- should be available and fully translated into. When it comes to chatbots, Rhode Island with Twilio gets top marks for availability in 5 languages (English, Spanish, Portuguese, Arabic, French).
This is much easier to do for Prompted and Support chat systems rather than free-text or live, because there is a known set of inputs from the customer (clicking on links or options) and outputs (pre-written responses). That doesn't mean multiple languages shouldn't be supported in Live and Free-text chat systems.
When choosing a vendor and building the contract, UI agencies should ask how the vendor will support them in providing help in multiple languages.
Mobile-friendly: For many people, their only way of accessing the internet is through their mobile phones, potentially with limited data plans. Chat integrations should be designed with this smaller screen in mind and be evaluated for how much data they use in the course of operating. If your integration is anything other than directly "out of the box," you should request support be made available to you for ensuring a great mobile experience.
Availability: You should make sure that the chat is available on any/all pages of your site (or sites!) that get meaningful traffic by people who might be looking for help with unemployment insurance. This would include any separate sites you might have for, e.g.: UI Agency administration, Standard UI/PEUC management, PUA management, re-employment services and support.
Specificity: In addition to wanting to put your chat content on multiple parts of your site or sites, you want to make sure that chat content from non-UI sites doesn't show up here. This shows up frequently in Support Chat implementations that draw on FAQ content from across the DOL. You should ask the vendor how to set up multiple instances of the chat
Data usage and speed: If the chat implementation is slow, it increases the odds of someone abandoning chat and turning to a different support channel. People are used to the internet being blazing fast and might find that more annoying and cumbersome than putting their phone on the table and doing other things while on hold. Part of chatbot speed is also related to how it is using data; this is also important as it relates to burdening people with limited data plans as little as possible.
Quality assurance:
It should be easy to test the chat before it becomes available on the website. This is especially true for live chats, to make sure that agents are comfortable with their side of the interface.
If the chat is available on multiple pages or sites, it should be easy to test it before it goes live on those different pages. Testing on multiple pages is important to ensure that the chat is always "on top" of the content of the rest of the page; you don't want page content hiding the chat