Real public-sector form

Utah Benefits

SNAP • Medical • Child Care • Financial Assistance

The Utah Department of Workforce Services provides a multi-program application that allows residents to apply for SNAP, financial assistance, child care and medical assistance.

Source: State of Utah Department of Workforce Services

AccessForce is an independent prototype. It is not affiliated with, nor endorsed by, this agency, and it never changes eligibility requirements or makes eligibility decisions.

View Official Utah Application Page Run the AI audit

Before

The experience people use today

31 pages.
Four assistance programs.
One application.

But not every applicant needs to answer every question.

Before answering the questions, applicants may first have to understand which questions they're supposed to answer.

Which questions apply to you?

The same application asks for different question ranges depending on which programs someone is requesting.

SNAPQuestions 1, 4–5, 8–25, 27–33+ SNAP section + signature
Financial AssistanceQuestions 1, 4–5, 7–10, 12–30+ Financial section + signature
Child CareQuestions 1, 4–5, 7–9, 12–23, 30+ Child Care section + signature

AI-identified cognitive and process complexity — not a legal accessibility violation.

Document preview

Representative excerpt

Application for Assistance — multi-program (form 61APP)

Cover page

Apply for: Food Stamps (SNAP) • Financial Assistance • Child Care • Medical Assistance
Complete only the questions that apply to the programs you are requesting.

Instruction block

If you are applying for SNAP, answer questions 1, 4–5, 8–25 and 27–33.
If you are applying for Financial Assistance, answer questions 1, 4–5, 7–10 and 12–30.
Sign and date the program section that applies to your request.

Numbered questions

12. List all household members, their relationship to you, and their date of birth.
19. Report all earned and unearned income for each household member.
27. Report resources, including bank accounts, vehicles and property.

Short representative excerpts only — the full published document is linked above.

AXS — Accessibility Experience Score

Putting a number on the experience people face today

AXS — AccessForce Accessibility Experience Score

42 / 100

Significant OpportunitiesIllustrative score until audit completed

Illustrative score until audit completed

31-page multi-program assistance application

AXS weighs technical accessibility at 40%, cognitive and content accessibility at 30%, and process accessibility at 30% — so an interface that is technically operable can still score low when the process itself is hard to get through.

How AXS works

AI accessibility audit

Ask AI to examine the experience

AccessForce reads the published application the way different people would have to, and reports what may stand in their way. These are AI-identified opportunities reviewed by a person — not legal findings.

AI redesign

Let Agentforce rebuild the interaction

The eligibility rules and the systems behind them stay exactly as they are. Only the interaction changes.

After

The same service, as a conversation

Utah Benefits — assistant

Hi! What can we help you with today?

You can choose more than one.

Before vs. after

Same public service, different experience

Before

31-page multi-program application

📄 Document-oriented
🔀 Program-specific question paths
🏛 Government terminology
👀 Significant visual scanning
❓ Applicant determines applicable sections

After

Adaptive Agentforce conversation

💬 Conversational
1️⃣ One question at a time
🧠 Plain language
🌎 Language choice first
♿ Accessibility preferences first
🤖 AI handles branching
📱 Mobile-first
🙋 Human help available

Same programs.

Same eligibility requirements.

A radically different experience.

Measured

Before and after isn't just visible. It's measurable.

Before

AXS / 100

Significant Opportunities

31-page multi-program assistance application

AccessForce redesign

After

AXS / 100

Highly Accessible Experience

Adaptive Agentforce conversation over the same requirements

Accessibility Experience Points

42 → 93. Illustrative score until audit completed.

Dimension by dimension

Select “Why?” on any dimension to see the finding, the recommendation and the redesign decisions the improvement is credited to.

AXS dimension scores before and after the AccessForce redesign, out of 100
Dimension	Before	After	Change
Visual & Perceivable	72 out of 100 before	95 out of 100 after	+23
Severity: LowImpact area: Visual & Perceivable Finding Dense pages ask people to scan long blocks of small print to find what applies to them. AccessForce recommendation Present one question at a time, at a size and contrast the person can adjust. Before 7 / 10 After redesign 10 / 10 Improvement +3 Why did the score change? One question on screen at a time Larger default text with an adjustable size control Information never carried by colour alone Layout works at 400% zoom and on a phone
Screen Reader & Semantic	60 out of 100 before	96 out of 100 after	+36
Severity: MediumImpact area: Screen Reader & Semantic Finding Document structure and field relationships are hard to follow when the experience is read aloud. AccessForce recommendation Rebuild the interaction with real headings, labelled controls and a predictable reading order. Before 6 / 10 After redesign 10 / 10 Improvement +4 Why did the score change? Real headings, labels and landmarks Every answer control has an accessible name Announcements when the assistant replies Reading order matches the visible order
Motor & Keyboard	57 out of 100 before	96 out of 100 after	+39
Severity: MediumImpact area: Motor & Keyboard Finding Completing the experience assumes precise input and, on paper, handwriting and signatures. AccessForce recommendation Make every step keyboard-operable with large targets, and accept spoken answers. Before 6 / 10 After redesign 10 / 10 Improvement +4 Why did the score change? Everything reachable with Tab, Enter and the arrow keys Visible focus on every control Large touch targets Nothing times out
Cognitive Accessibility	28 out of 100 before	91 out of 100 after	+63
Severity: HighImpact area: Cognitive Accessibility Finding Applicants must remember substantial historical information while navigating the application. AccessForce recommendation Break the process into smaller steps, allow approximate answers where appropriate, save unresolved information and return to it later. Before 3 / 10 After redesign 9 / 10 Improvement +6 Why did the score change? Reduced information density One question at a time Plain-language explanations “I don't know” supported Approximate answers supported where appropriate Follow-up items saved rather than blocking progress
Plain Language	32 out of 100 before	94 out of 100 after	+62
Severity: HighImpact area: Plain Language Finding Government and legal terminology assumes familiarity most applicants don't have. AccessForce recommendation Ask in plain language and explain terms in place, while preserving the original meaning. Before 3 / 10 After redesign 9 / 10 Improvement +6 Why did the score change? Questions rewritten at a grade 6–8 reading level Original wording available on request Acronyms and legal terms explained in place Underlying legal meaning preserved
Process Accessibility	21 out of 100 before	89 out of 100 after	+68
Severity: HighImpact area: Process Accessibility Finding Program-specific question ranges require applicants to map their own situation onto the structure of the form. AccessForce recommendation Ask what assistance the person needs, then let Agentforce decide which underlying questions apply. Before 2 / 10 After redesign 9 / 10 Improvement +7 Why did the score change? Irrelevant questions removed from view Branching handled dynamically Save and resume added Help available throughout One task presented at a time Human escalation available
Adaptive & Personalized	29 out of 100 before	94 out of 100 after	+65
Severity: HighImpact area: Adaptive & Personalized Finding The experience is the same for everyone, and assistance has to be discovered before it can be used. AccessForce recommendation Ask about accessibility needs and language first, then adapt the experience to the answers. Before 3 / 10 After redesign 9 / 10 Improvement +6 Why did the score change? Accessibility preferences asked before any application question Language choice offered up front Read-aloud and voice answers supported Simpler wording on request, at any point

Illustrative score until audit completed. Each change is credited to a specific redesign decision — a score never moves simply because a conversation exists.

Why did the score change?

Process Accessibility
21 → 89+68
Why?
- Irrelevant questions removed from view
- Branching handled dynamically
- Save and resume added
- Help available throughout
- One task presented at a time
- Human escalation available
Cognitive Accessibility
28 → 91+63
Why?
- Reduced information density
- One question at a time
- Plain-language explanations
- “I don't know” supported
- Approximate answers supported where appropriate
- Follow-up items saved rather than blocking progress

Scale

One accessibility engine. Thousands of experiences.

The same audit-and-redesign engine can point at any required experience. The systems of record behind them are never replaced.

Government
Healthcare
Education
Employment
Sports
Travel
Emergency Response
Financial Services

Don't make people adapt to forms.

Make forms adapt to people.

AccessForce: Forms for All

Built for Agentforce for Good • AI Fluency Track • Type C

See the Type C evidence How it reaches existing systems

Utah Benefits

The experience people use today

Which questions apply to you?

Document preview

Putting a number on the experience people face today

Ask AI to examine the experience

Let Agentforce rebuild the interaction

The same service, as a conversation

Same public service, different experience

31-page multi-program application

Adaptive Agentforce conversation

Before and after isn't just visible. It's measurable.

Dimension by dimension

Finding

AccessForce recommendation

Why did the score change?

Finding

AccessForce recommendation

Why did the score change?

Finding

AccessForce recommendation

Why did the score change?

Finding

AccessForce recommendation

Why did the score change?

Finding

AccessForce recommendation

Why did the score change?

Finding

AccessForce recommendation

Why did the score change?

Finding

AccessForce recommendation

Why did the score change?

Why did the score change?

Process Accessibility

Cognitive Accessibility

One accessibility engine. Thousands of experiences.