Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI agent for ux/ui designers

Heuristic Evaluation Sweep Agent

Deliver a deduplicated, severity-ranked usability issue list that is verified after fixes.

Quality controlCreativesProduct development
See how it worksGet access
Heuristic Evaluation Sweep Agent: what goes in, what the agent does and what you get

What it does

A review before launch usually catches whatever the reviewer happens to click on. This agent walks through the main flows of a prototype or live product in a set order, such as sign-up, search and checkout. At each screen it applies a heuristic checklist, covering visibility of status, error messages, consistency, recovery from mistakes and clarity of labels. It records each issue with the screen, a screenshot, a severity from 1 to 4 and the heuristic broken. It then groups duplicates so the same bad label on ten screens becomes one issue. After the team ships fixes, it walks the same flows again and checks that each fixed issue is gone and that no new issue appeared. The designer approves the final issue list before it goes to the team. Edge case: a screen only fails on mobile, so the agent repeats that flow at phone width.

How it works

Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
ApprovedYes, continueYes, continueNoNo 1 STARTS WHEN A flow is ready for review 2 USES A TOOL Open the product and walk each main flow in order 3 DOES Apply the heuristic checklist on every screen 4 USES A TOOL Capture screenshots and record each issue 5 DOES Group duplicates and set severity 6 YOU APPROVE Designer approves the issue list for the team 7 USES A TOOL After fixes ship, walk the same flows again 8 CHECKS THE RESULT Is every fixed issue gone from the screen? If not: reopen the issue with a new screenshot and rerunthe checklist on that screen. Back to step 3. 9 CHECKS THE RESULT Did the fixes cause any new issue? If not: log the new issue and walk the nearby screensagain. Back to step 3. 10 RESULT Verified issue list with open items
Read the steps as a list
  1. A flow is ready for review
  2. Open the product and walk each main flow in order
  3. Apply the heuristic checklist on every screen
  4. Capture screenshots and record each issue
  5. Group duplicates and set severity
  6. Designer approves the issue list for the teamThe agent waits here for your OK.
  7. After fixes ship, walk the same flows again
  8. Is every fixed issue gone from the screen?If not: reopen the issue with a new screenshot and rerun the checklist on that screen. Back to step 3.
  9. Did the fixes cause any new issue?If not: log the new issue and walk the nearby screens again. Back to step 3.
  10. Verified issue list with open items

How it decides

It marks a screen as failing a heuristic when a checklist question is answered no. It sets severity by how many users hit it and whether they can recover.

  • Severity 4 if users cannot finish the task, 3 if they finish with real difficulty, 2 if minor, 1 if cosmetic
  • Merge issues with the same cause on different screens into one
  • Repeat any flow at phone width when the product has a mobile layout
  • Reopen an issue when the fix only works in one browser

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Heuristic set to use (default ten usability heuristics)
  • Flows to walk and their order
  • Severity scale and cutoffs
  • Screen widths to test (default desktop and phone)
  • Where the issue list is written

What keeps you in control

It always asks you first

  • Designer approves the final issue list before it is shared

Hard limits

  • Never change the product or submit real orders
  • Never close an issue without checking the screen again

It stops when

  • Done: all severity 3 and 4 issues are verified fixed
  • Stop: the product needs a login the agent does not have

Set it up

We guide you through the set-up, step by step

Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
  • The agent then walks you through connecting your own data, one source at a time
  • A downloadable copy with the flow chart, the rules and the full guide
Get access to this agent

An example run

What happensThe agent walks a booking flow on 14 screens and records 31 issues. After grouping, 19 remain, including 3 at severity 4, such as a payment error that clears the form. The designer approves the list. After the fix release the agent walks the flow again. Two severity 4 issues are fixed, but the payment error still clears the card field on mobile, so it reopens that issue with a new screenshot.

More agents for ux/ui designers

Quality control · Creatives

Design System Drift Repair Agent

Working design files match the approved design library, without breaking layouts or intentional exceptions.

10 steps2 checks5 jobs
Quality control · Coding

Public Form Accessibility Agent

Every public form remains accessible after each revision.

7 steps2 checks5 jobs
Creating · Quality control

Prototype Interaction State Agent

Prototypes that cover errors and recovery, not just the happy path.

8 steps2 checks3 jobs
Quality control · Creatives

Accessibility Option Audit Agent

A tested accessibility checklist and an honest statement of what the game supports

10 steps2 checks4 jobs