Out Of AI℠ — All Things AI
Start here ↗
← All 45 workflows

Workflow 13 / Customer experience

Turn customer feedback into actionable themes

Group comments into traceable themes while keeping uncommon but important concerns visible.

What you need

An anonymized feedback sample with response IDs, collection dates and context about how responses were gathered.

What you’ll produce

A theme table, supporting examples and proposed service improvements for team review.

Follow the workflow

  1. Remove identifiers and choose a defined time period. Keep the sampling method visible so readers understand who is represented.
  2. Ask AI to propose themes and assign response IDs, allowing multiple themes where appropriate. Keep a separate list of unusual or serious concerns.
  3. Check a sample of assignments and count responses with spreadsheet or analysis tools. State whether counts are unique responses or theme mentions.
  4. Have the team select a small improvement to test and review subsequent feedback using the same collection method.

Your starter prompt

Group these anonymized comments into themes. For each theme provide a definition, response IDs, short supporting excerpts and a possible service improvement. Keep minority viewpoints and serious concerns in a separate section. Do not infer customer demographics or invent quotations. State when a comment fits multiple themes. Propose a counting method but do not claim the sample represents all customers. Sampling context: [paste] Comments with IDs: [paste]

Before using the result

Verify quotations and counts against the original comments. Human reviewers should handle serious complaints and avoid interpreting sentiment labels as facts about a person.

Measure whether it helps: Track reviewer agreement on themes, corrections, follow-up actions and feedback after a documented change.

Choose an approved tool

Start with the assistant your organization already approves. Work from a small sample and use only information you are permitted to share. File support and connected-app features depend on the product and plan. These guides do not connect to your accounts or run tasks for you.

Further learning: Official Gemini workplace learning. This is an original OutOfAI℠ workflow and prompt, not a vendor-certified procedure or a tested claim of time savings. Prepared October 22, 2026.

Choose another workflow →

Original training example · synthetic data

Illustrative feedback grouping

Information supplied

Twenty fictional comments: 11 mention late deliveries, 4 report damaged packaging, 3 request status updates and 2 describe accessibility barriers.

Acceptable draft

Group by theme while keeping the small accessibility group visible; do not discard issues based on frequency alone.

Required human decision

Customer experience owner checks source comments and appropriate escalation.

For the full example, controls and measurement approach: Read the OutOfAI worked guide →.