As of 13 August 2026, AI can create customer personas from your research data.

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 1 hour until you can act on the result.

   How to actually do it

   1. Open the research repository, survey export and product-usage reports, then remove names, email addresses, account numbers and other direct identifiers before copying anything.
   2. Write down the product, UK market, research question and decision the personas must support.
   3. Paste the anonymised material and that context into a chatbot using the supplied prompt, keeping interview, survey and behavioural sources clearly labelled.
   4. Ask the model to add a source reference to each finding, then compare every persona claim with the original note, response or data row.
   5. Merge duplicated personas and challenge any group based only on age, gender, location or other demographic assumptions unless the research shows a meaningful behavioural difference.
   6. Ask a product or research colleague to review the evidence, confidence levels, missing data and proposed recommendations before the personas are shared.
   7. Store the approved personas with their evidence and limitations, and record which product or research decision each persona is allowed to inform.

   Prompt

   Create evidence-based customer personas from the research data below.
   
   Context:
   - Product or service: [describe it]
   - Market and country: [for example, UK]
   - Research question: [what you were trying to learn]
   - Intended decision: [what the personas will be used for]
   
   Research data:
   [Paste anonymised interview notes, survey responses, behavioural data and relevant customer context here. Remove names, email addresses, account numbers and other direct identifiers.]
   
   Instructions:
   1. Use only the supplied data. Do not invent demographics, motivations, quotes, behaviours or sample sizes.
   2. Separate direct evidence from interpretation. Mark interpretations as hypotheses.
   3. Group customers by meaningful differences in needs, behaviours, constraints or goals, not by stereotypes or arbitrary demographic labels.
   4. Say when the evidence is too weak to support a separate persona or a claim.
   5. Create no more than five personas unless the data clearly supports more.
   6. For each persona provide: a neutral name, one-sentence summary, supported goals, needs, behaviours, barriers, relevant context, evidence from the data, confidence level, and unanswered questions.
   7. Include the number or proportion of research participants only when it is present in the supplied data.
   8. Do not use verbatim quotes unless they appear in the source data, and label each quote with its source reference if one is provided.
   9. Add a comparison table showing how the personas differ and identify any customers who do not fit the proposed groups.
   10. Finish with three practical research or product decisions these personas could inform, clearly labelled as recommendations rather than findings.
   11. List possible sampling bias, missing evidence and privacy risks before the final output is used.
   
   Return the result in clear Markdown with headings and tables.

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