AI Flow Finder
Data workflow

Group feedback into a traceable theme map

Organize a small feedback sample by theme while keeping each comment tied to its original ID.

◦ Documented~20 minutes · EstimateBeginner

Start with these inputs

  • Anonymized feedback with stable comment IDs
  • A small starting list of themes, or permission to propose one
  • Rules for multiple themes and uncertain classifications

Follow the workflow

  1. Remove names and sensitive details, assign comment IDs, and keep a separate source copy.
  2. Paste a manageable sample into Claude with your labeling rules.
  3. Review the theme glossary and spot-check every uncertain or multi-label comment.
  4. Recount the reviewed labels in a spreadsheet and keep the ID mapping with the summary.

Your starting prompt

PERSONALIZE THE [BRACKETS] BEFORE USING
Classify the feedback below into a small set of themes.
Return: a proposed theme glossary; a table of comment ID, theme, supporting phrase, and uncertainty; and a summary of patterns. Keep every original ID. Allow "Unclear" and multiple labels where justified. Do not invent comments or infer personal traits.
If you report counts, show which IDs support each count and explain whether comments can appear in multiple themes. Do not describe this sample as representative of all customers.
Starting themes/rules: [paste]
Feedback: [paste ID-labeled comments]

Three workflows. One useful email.

Practical guides, copyable prompts, and a short review tip. Every Tuesday. Confirm your email to join; unsubscribe anytime.

Get the weekly workflows

What the output could look like

Illustrative example, not a recorded test result.

Illustrative example using invented sample inputs: Theme: onboarding clarity. Supporting IDs: C01, C04, C07. Uncertain: C04 may also describe a missing feature. Pattern: several comments ask what to do first; this small sample does not establish overall demand.
Review before you use it.
AI classifications and counts need review. Qualitative themes are interpretations, and a convenience sample cannot establish population-level conclusions. Time is an approximate planning estimate, not a measured result. Free access has usage limits; features can vary by account.

Sources & review notes

By AI Flow Finder · Reviewed October 14, 2026. Documented means this guide is based on the sources above. It has not yet passed a recorded end-to-end workflow test. Tool access and limits can change.

Pass on a useful workflow

Keep the momentum going