Screenshot to UX Audit

Running it

1. 1

   Supply context first

   Who the users are, what the primary task on this screen is, and what constraints apply. Without this the model reviews a generic interface.

2. 2

   Name the framework

   Ask for a review against Nielsen's ten heuristics, or against a specific list, so the output is structured and checkable.

3. 3

   Ask for evidence

   Require each issue to name the element, the heuristic and the likely consequence, so you can verify it.

4. 4

   Ask for severity and rationale

   Frequency, impact and persistence, with the reasoning stated.

5. 5

   Verify every item

   Check each finding against the real interface before it enters a report.

A prompt that produces checkable output

Review this screen against Nielsen's ten usability heuristics.

Context:
- Product: [what it is]
- Primary user: [who, and their expertise]
- Primary task on this screen: [the one thing]
- Constraints: [platform, regulatory, technical]

For each issue return:
1. Heuristic violated
2. The specific element, described so I can find it
3. What a user would experience
4. Severity (critical / high / medium / low) with reasoning
5. A concrete change, not a general principle

Do not list issues you cannot see evidence for in the image.
If the screenshot does not show enough to judge something, say so.

Where it genuinely helps

• Catching obvious contrast, hierarchy and label problems quickly across many screens.
• Producing a structured first draft that a reviewer edits rather than writes.
• Reviewing your own work, where the curse of knowledge is strongest.
• Generating a checklist of things to verify manually.