Author: AILesson10 min setupTested with:ChatGPTReviewed: 2026-08-28
Quick answer
Cluster feedback by underlying progress and breakdown while preserving source, variation, and uncertainty. Provide: Feedback corpus, Sample and quality context, Product decision context. Expected result: An evidence-linked opportunity map with themes, affected contexts, negative cases, confidence, risks, and next research.
1
Add your context
Your text stays in this browser. AILesson Prompts does not send it to a model or server.
2
Your prompt
Unfilled fields remain visible as placeholders, so you can still copy and edit the prompt
Organize the supplied user feedback into evidence-bound opportunity areas, not a feature backlog.
Feedback records, IDs, dates, channels, roles, context, wording, versions, outcomes, and consent:
[feedback]
Window, bias, duplicates, exposure, missingness, taxonomy, prior themes, and incidents:
[quality]
Horizon, strategy, decisions, constraints, risk, priority criteria, excluded solutions, and reviewers:
[decision]
Validate unique records and distinguish unsolicited feedback, prompted answers, observed behavior, support issue, feature request, workaround, satisfaction, and researcher interpretation. Code situations, triggers, goals, current alternatives, breakdowns, consequences, constraints, and desired progress with source locators. Cluster by solution-independent opportunity, not shared keywords or requested features. For each area write a “people in context need progress because evidence” statement; list unique records and participants, roles and segments, supporting behaviors and quotes, negative and contradictory cases, known-incident overlap, severity and frequency evidence, confidence, strategic relevance, risks, and unknowns. Do not infer prevalence without a denominator, combine incompatible contexts, count repeat contacts as new users, or call an idea validated because people requested it. Rank only with supplied criteria and show sensitivity; otherwise provide a review order. Separate opportunity evidence from solution ideas and recommend the next smallest research or instrumentation step.
Private by defaultPrompt assembly happens locally in your browser. Avoid placing confidential information into any AI service unless your organization allows it.
From input to outcome
A worked example
See how concrete context turns this recipe into a usable result
Actual input
Feedback corpus
Last 60 days: 42 unique support cases tagged reminder; 18 unique cases request fewer notifications; 8 interview participants from support recruitment. P01 and P03 could not identify an accepter in recent handoffs; P02 intentionally waited across timezone and disliked more email; P04 waited for known approval. Analytics: 31% of active workspaces had one Ready state over 48 hours, but event cannot distinguish intentional waiting. Incident INC-901 caused duplicate email for 11 of the 18 fewer-notification cases; fixed Aug 12. Record IDs and locators are available; one support customer contacted three times but is one case.
Sample and quality context
Support channels overrepresent problems; interviews are a convenience sample. Deduplicate by case and participant. Active-workspace denominator is known for analytics only. Versions missing in 20% of support cases. Feature-request prompts were used in two interviews. No causal link between long Ready and missed work.
Product decision context
Next-quarter goal is reduce avoidable handoff uncertainty without increasing notification burden. Options remain open; automatic email is not preselected. Review order weighs severity, strategic fit, evidence strength, and reach, but no numeric weights approved. Privacy and accessibility review required for concepts. Product and Research review together.
Example output
Opportunity 1 — make acceptance responsibility and next authority legible in ambiguous handoffs
Statement: when a handoff is ready but acceptance responsibility is not visible, acting roles need to identify who can accept and what happens next so they can route action without searching across people. Evidence: P01 and P03 recent behaviors and artifacts; medium confidence, two convenience-sample participants. Negative/variant cases: P02 intentionally waited and P04 knew approval was pending, so not every long Ready state belongs here. Strategic fit is high; reach is unknown. Next step: instrument accepter presence and recruit recent instances across support and non-support populations.
Opportunity 2 — distinguish expected waiting from unattended work
Statement: when Ready persists, roles need the reason and expected next event so they can decide whether to wait or intervene. Evidence: P02 timezone wait, P04 approval wait, plus analytics that detect duration but not reason. Medium confidence on variation, low on reach and consequence. Next step: taxonomy study and event-definition audit.
Opportunity 3 — control notification relevance and duplication
Evidence: 18 unique fewer-notification cases, but 11 overlap fixed duplicate-email incident INC-901. Post-fix residual is seven cases, with version missing in some records. Do not merge all 42 reminder-tag cases as support for more email; tags may represent requests, current failures, or discussion. Incident evidence is separate from the broader opportunity. Next step: recode remaining cases by trigger, channel, desired action, and post-fix version.
Review order: first opportunities 1 and 2 because they align with reducing uncertainty without selecting a notification solution; then assess residual notification control after incident cleanup. Numeric ranking is not justified without approved weights and comparable reach. The 31% workspace metric describes an event condition only, not avoidable delay or affected-user prevalence. Automatic email remains one unvalidated idea and requires privacy/accessibility review.
Why this works
1
Solution-independent clustering preserves the problem space before feature selection.
2
Negative cases and channel bias keep noisy feedback volume from becoming false certainty.
Check the result
Does each opportunity combine compatible contexts and trace to unique evidence?
Are requests, incidents, behavior, interpretation, and solution ideas separate?
Are prevalence, priority, and confidence proportional to sample quality and criteria?
Use it with confidence
Frequently asked questions
Practical answers about when to use this recipe, what to provide, and where human review still matters
What should I prepare before using “Organize user feedback into opportunity areas”?
For “Organize user feedback into opportunity areas,” prepare Feedback corpus, Sample and quality context, and Product decision context. Replace placeholders only with information you can verify. If a detail is unknown, preserve that uncertainty explicitly instead of asking the model to infer it.
When is the “Organize user feedback into opportunity areas” result not ready to use?
The result is not ready if it does not yet deliver the stated outcome—An evidence-linked opportunity map with themes, affected contexts, negative cases, confidence, risks, and next research—from the supplied evidence, or if it relies on unresolved assumptions, missing approvals, or invented details. Use the checks as release gates: revise the source inputs or assign a named, authorized reviewer instead of polishing an unsupported output.
Which AI tools have recorded tests for “Organize user feedback into opportunity areas”?
The published test record for “Organize user feedback into opportunity areas” lists ChatGPT as of 2026-08-28. This confirms recorded runs, not guaranteed compatibility or identical results in later product versions. For another tool or version, keep every constraint visible and repeat the result checks before use.
ChatGPT