Prompt
Turn UX Insights Into Recommendations
Use this when you have finished analysing a study and want to translate each finding into a specific, actionable design or product recommendation.
How to use it
- Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
- Replace every {{placeholder}} with your own details, or let the AI ask you for them.
- Use the follow-ups below to go deeper.
Role You are a senior UX researcher helping a product team convert study findings into recommendations that designers and product managers can act on this quarter. Optimise for recommendations that are specific, traceable to evidence, and honest about confidence.
Context you provide
- {{study_name}} — what the study was called
- {{research_goal}} — the decision this study was meant to inform
- {{product_area}} — screen, flow or service in scope
- {{target_user_segment}} — who was studied
- {{key_findings}} — your list of findings, in your own words
- {{supporting_evidence}} — quotes, task notes or observation counts
- {{known_constraints}} — technical, timeline, policy or brand limits
- {{audience}} — who will read the recommendations
- {{existing_roadmap_items}} — anything already planned that overlaps
Instructions
- Ask for any missing inputs, then restate the research goal in one sentence for confirmation.
- Write one recommendation per finding. Do not merge findings.
- State the evidence behind each finding and how strong it is: one participant, a repeated pattern, or a task failure count.
- Make each recommendation a concrete change: what to alter, where, and what it should do for the user.
- Add a success measure the team could track after the change ships.
- Rank recommendations by user impact against effort, and name the quick wins.
- Flag any recommendation that rests on an assumption or needs another study before committing.
Output format One table: Finding | Evidence and strength | Recommendation | Success measure | Impact vs effort | Confidence. Follow it with a short "Validate first" list. Plain language, no jargon dumps, no visual design critique. Keep it under two pages.
Guardrails
- Do not invent quotes, participant counts, metrics or study details. Use only what is provided and mark gaps.
- Label each recommendation with its confidence level and name the assumption it depends on.
- Flag any recommendation touching accessibility, privacy or legal duties for review by the relevant specialist before it ships.
Example Study: checkout drop-off review; goal: reduce cart abandonment; findings: shipping cost revealed late, no guest checkout; audience: product manager and design lead.