Prompts / feedback_synthesis
Synthesise feedback into themes and decisions Free
Turns a pile of feedback (interviews, tickets, survey comments, critique notes) into evidence-backed themes, severity and next actions.
Run it
- Claude Code:
/mcp__designers__feedback_synthesis - Claude Desktop and claude.ai: plus menu, Add from Designers, Synthesise feedback into themes and decisions
- Any app: Use the feedback_synthesis prompt from the Designers server on the raw feedback, pasted or described.
Arguments
| Name | Meaning |
|---|---|
feedback * | The raw feedback, pasted or described |
goal | The question you are trying to answer |
* required
What your assistant is told to do
The full instructions, so you can see exactly what you are asking for:
Synthesise this feedback. Work bottom-up: do not start with the themes you expect.
Goal: <goal>
Feedback:
<feedback>
Method:
1. Split into single observations, one idea each, keeping the person's own words. Tag each with who said it and the context, if known.
2. If the feedback is design critique from colleagues, call lint_design_feedback and mark which notes are ready to act on; set aside opinions and solutions-without-problems, but keep the underlying problem if one can be recovered.
3. Cluster by theme. Name each theme as a finding in a full sentence ("People do not notice the save state after editing"). Count how many sources support each, and note how strongly each person felt.
4. Separate what was observed from what you infer. Mark where the evidence is thin or comes from one person.
5. Rate each theme for impact on the goal and frequency; put the cost of getting it wrong in the severity.
6. Output: a table of themes (finding, evidence count, example quotes, severity, suggested action), then a short list of decisions to make this week, and questions the data cannot answer plus the cheapest way to answer each.
Do not average opinions. Disagreement between people is a finding.