Skip to main content

Beta: AI insights into your manual evaluations

  • Updated

This feature is currently in beta. To learn more about joining the beta program, reach out to our team at beta@levelaccess.com.

You can use built-in AI prompts and get concise, AI powered summaries of your manual evaluations. They help you to quickly identify what to fix first. Use them to move directly from interpretation of detected issues to their remediation.  Focus on the most important issues from your manual evaluation, and assess and prioritize them quickly and efficiently. 

On this page: 

Use AI prompts to summarize results of a manual evaluation

To access and use AI prompts at the Insights tab:

  1. Go to Digital assets.
  2. Select Evaluations.
  3. Select a manual evaluation you want to learn more about.
  4. Choose the Insights tab.
  5. Select the AI prompt that best matches your immediate goal. A generated response summarizes the key points you need to consider. Refer to What's in a generated AI response? 
  6. To get a new version, at the end of the response, select Regenerate response.
  7. Initiate the remediation workflow in the platform. For example, click an issue, view the guidance, and create tasks.

Which AI prompts are available at the Insights tab?

Each AI prompt is designed to help you interpret the manual evaluation so that you can prioritize findings and start remediation. When you ask a question, using one of the prompts, the platform:

  • Generates a plain-language summary tailored to the selected prompt
  • Surfaces the most impactful issues with contextual examples
  • Links directly to relevant findings so you can explore details and take action

All AI generated responses are saved per evaluation and prompt, and have a timestamp. You can navigate between different responses to your prompt, by clicking the Previous and Next arrows, at the top of the response. 

This table lists the AI prompts at the Insights tab, and explains what happens when you use them.  

When to use this prompt What happens? 
What disabilities are most impacted?

Groups issues by disability type. Explains how each set of findings affects users of your digital asset. Use this prompt to highlight disability groups that are experiencing the most barriers when using the evaluated digital asset. 
Which screens are the least accessible? Highlights specific screens with most issues. Use this prompt to identify the least accessible part of your product. Findings are listed under the key screens and provide remediation recommendations.
How should I prioritize if I’m optimizing for risk? Organizes findings by severity, affected disability types, and age. Use this prompt to identify issues of the highest priority, because they have the highest impact or exposure.
How should I prioritize if I’m optimizing for a VPAT?

Groups findings by WCAG success criteria to support teams working on accessibility reporting or documentation. Explains how users are affected.

If the evaluation finding does not contain an accessibility standard, you must select one before the AI response can be generated. If you select WCAG, you must also provide a version and conformance level. For example, WCAG 2.2 AA.

What's in a generated AI response?

Depending on the AI prompt you used, an AI generated response might include:

  • A summary of the detected issues.
  • Priority ranking, with most important groups of issues listed first.
  • Links to the example issues, the referenced issues, and the filtered findings view, which allows you to explore the issues further. For example: "View findings for success criteria: 2.1.1."
  • User impact, based on a particular WCAG standard criterion.
  • Recommendation on the next steps to take to remediate the issue.
    • For example, use the following prompt: How should I prioritize if I’m optimizing for a VPAT?
    • The AI generated response might highlight the following guideline: WCAG 2.1.1 - Keyboard.
    • The AI response then provides the following recommendation: "Users with mobility disabilities who rely on keyboard navigation cannot access functionality that’s only available through mouse interactions."

FAQs about AI Insights prompts and responses 

Read responses to some frequently asked questions about built-in AI prompts: 

  • Why do I see a banner asking me to select standards?
    Some evaluations don’t include standards in the uploaded evaluation file, depending on the format or tool that was used to upload them to the platform. To generate a VPAT focused summary, the platform needs to know which standard you're targeting. 
  • Do I have to select standards every time?
    No. After they are saved, standards apply to future VPAT summaries for that evaluation. You’ll be prompted again only if you regenerate the AI response and there’s no cached version. 
  • Why doesn’t my summary include all the findings from the audit?
    AI insights include only unresolved and actionable findings. They exclude findings with these statuses: Dismissed, Fixed, Cannot be fixed, Not reproducible, To review. This reduces clutter and focusses on already fixed findings.
  • Can I use AI summaries to start remediation? Yes. Summaries link directly to findings. You can navigate to any finding, view details, and create tasks.
  • What happens when I click on a finding link in the AI summary?
    You’ll be taken to the full finding detail in the evaluation where you can take further action - review it, assign it, resolve or dismiss it.
  • Can I use this to decide what to fix first? 
    Yes. Each prompt provides contextual guidance to help you prioritize your remediation work.

Related articles

Comments

0 comments

Please sign in to leave a comment.