As of 13 August 2026, AI can prioritise which customer feedback to act on first.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open your survey, support, review and complaint sources, export the relevant feedback with its date, channel, product area, customer segment and any available case or ticket identifier, then remove direct personal details that are not needed.
   2. Write down the business goals, available team capacity, known product constraints and any issues that must receive urgent human attention, such as privacy, security, safety or legal complaints.
   3. Paste the business context and feedback into the prompt, keeping the original comment text or a traceable identifier beside each item so the proposed themes can be checked.
   4. Ask the chatbot to produce the ranked table, including separate scores for impact, reach, urgency, confidence and effort, rather than asking for a single unexplained priority score.
   5. Compare every high-priority theme with the original comments and your support or product records, correcting duplicate counts, missing context and any theme that the model has combined incorrectly.
   6. Send urgent, sensitive or ambiguous items to the relevant support, product, security or legal owner, then choose the actions and owners for the remaining priorities.
   7. Record the chosen actions and revisit the ranking after new feedback or outcome data arrives, rather than treating the first AI ranking as a permanent decision.

   Prompt

   Prioritise the customer feedback below into a ranked action list for a UK business.
   
   Business context:
   - Business and product: [describe them]
   - Main customer segment: [describe it]
   - Current business goals: [list them]
   - Constraints, such as team capacity, technical limits or contractual commitments: [list them]
   - Time period covered by the feedback: [state it]
   
   Use this method:
   1. Remove duplicate comments but retain the number of similar comments and show when a theme is based on only one comment.
   2. Group comments into clear themes and quote or reference the source comments for each theme.
   3. Score each theme from 1 to 5 for customer impact, number of customers affected, urgency, confidence in the evidence and effort to address. Do not invent scores from missing information. Mark missing information as unknown.
   4. Rank the themes using customer impact, reach and urgency first, while showing confidence and estimated effort separately rather than hiding them in one unexplained score.
   5. Flag complaints involving safety, discrimination, privacy, security, legal issues, vulnerable customers or possible regulatory breaches for immediate human review instead of treating them as ordinary product requests.
   6. Recommend the next action for each theme, such as investigate, reply to affected customers, fix a defect, update guidance, monitor or reject as out of scope.
   7. Give me the top priorities in a table with: rank, theme, evidence and count, likely customer impact, urgency, confidence, effort, recommended next action and what I still need to check.
   8. State clearly which conclusions are uncertain. Do not invent customer numbers, causes, revenue impact, deadlines or facts not present in the material.
   
   After the table, give me three questions I should answer before committing resources. Do not make the final business decision for me.
   
   Feedback:
   [paste the feedback export or comments here]

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.