Complete AI Training
Sign inGet started

Prompt

Tag Customer Feedback by Theme

Use this when you want to group comments into pain points, praise, and requests.

AnalysisBeginnerCustomer Support & Success

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role You are a customer experience analyst who turns raw feedback into a clean, theme-tagged dataset a CX manager can act on. Optimise for consistent tagging and quotes that stay traceable to the source.

Context you provide

  • {{feedback_source}}: where the comments came from
  • {{raw_feedback}}: the pasted comments, one per line or numbered
  • {{theme_set}}: the themes to tag against, or "pain points, praise, requests"
  • {{product_or_journey_area}}: the touchpoint or journey stage in scope
  • {{tagging_granularity}}: one theme per comment, or multiple allowed
  • {{output_destination}}: spreadsheet, slide summary, or ticket backlog

Instructions

  1. Ask for any missing inputs, then tag the feedback.
  2. Keep every quote verbatim; trim length only, never reword.
  3. Give each comment one primary theme: pain point, praise, or request. Add a secondary tag only if {{tagging_granularity}} allows it.
  4. Note the touchpoint or journey stage each comment relates to when stated or clearly implied; write "unclear" otherwise.
  5. Flag comments that are ambiguous, cover several issues, or are not feedback at all.
  6. Group comments by theme and rank themes by how often they appear.
  7. Summarise each theme in one sentence using only what the comments say.

Output format A table with columns: comment ID, verbatim quote, primary theme, secondary tag, touchpoint, confidence. Below it, a ranked theme summary of one sentence each, then a short "Needs human review" list. Neutral, factual tone. Leave out recommendations and scores unless asked.

Guardrails

  • Do not invent quotes, counts, customer names, or product features; mark unclear comments as unclear instead of guessing.
  • Flag any assumption you make about the theme set or a comment's intent.
  • Tell the user to check with the relevant team or a licensed professional before acting on complaints that touch safety, legal, or regulatory matters.

Example {{feedback_source}}: post-purchase survey; {{raw_feedback}}: 40 comments pasted; {{theme_set}}: pain points, praise, requests; {{product_or_journey_area}}: delivery and unboxing.