As of 13 August 2026, AI can audit your website UX.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, chat-fluent skill, and roughly 1 hour until you can act on the result.

   How to actually do it

   1. Open the key pages and user journeys in your website, then record the starting page, intended task and point where a user is expected to finish.
   2. Capture screenshots or a screen recording of the important desktop and mobile states, including menus, forms, validation messages and any error or confirmation pages.
   3. Gather the target user description, business goal, known constraints, analytics exports and relevant user research, removing private customer data before sharing it.
   4. Paste the prompt into a chatbot, attach the screenshots and data, and replace each bracketed slot with the context for this website.
   5. Compare every reported issue with the corresponding page or data, mark findings that are only assumptions, and remove recommendations that have no supporting evidence.
   6. Run the proposed tests with representative users or existing analytics, then turn the confirmed findings into a prioritised backlog for your product or development team.

   Prompt

   Audit the UX of this website using only the material I provide.
   
   Website or page URLs: [PASTE URLS]
   Screenshots or screen recordings: [ATTACH THEM]
   Primary user: [DESCRIBE THE TARGET USER]
   Main task the user should complete: [DESCRIBE THE TASK]
   Business goal: [DESCRIBE THE GOAL]
   Known constraints: [TECHNICAL, LEGAL, BRAND OR DELIVERY CONSTRAINTS]
   Analytics, search data or user research: [PASTE OR ATTACH IF AVAILABLE]
   
   Produce the audit in this structure:
   1. Executive summary with the three most important findings.
   2. User journey review from entry point to completion.
   3. Findings table with location, observed evidence, affected user or task, severity, confidence, likely cause and recommendation.
   4. Separate direct observations from inferences and assumptions. Do not present an inference as evidence.
   5. Check content clarity, navigation, information hierarchy, forms, calls to action, error handling, mobile layout, keyboard access, colour contrast and obvious accessibility risks where the supplied material permits it. State when a check cannot be made from the material.
   6. Prioritise findings by user impact, reach, confidence and implementation effort. Explain each priority rather than inventing scores or conversion claims.
   7. Give a short test plan for the highest-priority findings, including the task to test, what to measure and what result would change the recommendation.
   8. Do not invent users, analytics, standards compliance, technical behaviour or business context. Do not claim that a change will improve conversion unless the supplied evidence supports that claim.
   
   Finish with a list of information or tests still needed before making a major design change.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.