As of 13 August 2026, AI can analyse your customer survey results.

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 the survey platform or spreadsheet and export the responses with the full question wording, response options, timestamps if relevant, and any legitimate segment fields.
   2. Remove names, email addresses, free-text personal details and other unnecessary identifying information, then record how the responses were collected, who was invited and the total number invited if known.
   3. Paste the survey objective, business decision, collection method, known limitations, full question list and anonymised data into the prompt.
   4. Ask the chatbot to produce the report, then ask it to show the calculation for every important count, percentage and segment comparison.
   5. Compare each reported count, percentage and quotation with the source spreadsheet, and correct any denominator, transcription or duplicate-response error before using the findings.
   6. Mark conclusions about bias, representativeness, causation and small groups as limitations unless your research method supports them, then send the checked report to a colleague with the proposed actions and tests.

   Prompt

   Analyse the customer survey data below for a UK business.
   
   Business context: [what we sell, who answered, and the decision this analysis will inform]
   Survey objective: [what the survey was intended to find out]
   Collection method and dates: [how and when responses were collected]
   Target population: [which customers or prospects the survey was meant to represent]
   Known limitations: [missing responses, incentives, duplicate responses, unusual recruitment, or other concerns]
   
   Survey questions and response options:
   [PASTE THE FULL QUESTION LIST]
   
   Survey data:
   [PASTE AN ANONYMISED CSV OR TABLE, OR DESCRIBE THE COLUMNS]
   
   Produce a concise report with these sections:
   1. Data checks: number of usable responses, missing values, duplicate or inconsistent entries, and any denominator used for each result. Do not invent a response rate when the total invited population is not supplied.
   2. Question-by-question results: counts and percentages where the denominator is clear, with the calculation shown for important figures.
   3. Open-text themes: group similar answers, state how many responses support each theme, include short representative quotations, and keep unusual but relevant responses visible.
   4. Segments: compare groups only where the data contains a clear segment field and the group sizes are stated. Flag small or uneven groups instead of presenting them as reliable differences.
   5. Interpretation: separate what the data directly shows from possible explanations. Do not claim causation, customer-wide representation or statistical significance unless the supplied data and method support it.
   6. Recommendations: give three actions ranked by likely usefulness, with the evidence for each, the key assumption and a low-cost way to test it.
   7. Limits and questions: list what the survey cannot establish and what extra information would change the conclusion.
   
   Use plain British English. Be precise and sceptical. Do not fabricate figures, quotations, respondent characteristics, statistical tests or business context. If a calculation or conclusion cannot be supported by the supplied data, say so. End with a short executive summary for a colleague.

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