A customer persona is a research-backed sketch of a representative buyer: who they are, what they are trying to achieve, what frustrates them, and how they behave. Good personas align product, marketing, and sales around the same human. The failure mode is inventing a persona from imagination, which produces a confident-sounding fiction that misleads every decision downstream. This builder turns your real inputs into a structured prompt that enriches a grounded draft and separates fact from assumption.
How it works
You start by describing the product type, the target market, and a short summary of any customer data you already hold — survey results, support themes, churn reasons, sales-call notes. You then fill in the persona scaffold: demographics, top goals, pain points, and observed behaviours. The tool weaves these into a prompt that instructs the model to enrich each section, propose a memorable name and one-line summary, map the persona to a jobs-to-be-done statement, and — crucially — label every enrichment as either grounded in your data or an assumption to validate. The prompt is generated locally; nothing is uploaded.
What a well-formed persona contains
A persona that drives real decisions needs more than demographics. The most actionable sections are:
The job-to-be-done — what the customer is trying to accomplish, framed as progress they want to make. “Help me file tax returns without errors” is more useful than “wants an accounting product” because it names the outcome.
The trigger — what event or situation causes them to start looking for a solution. Understanding the trigger tells you where and when to reach them.
The barrier — what stops them from buying or switching. This is often more important than the job itself; if you cannot address the barrier, nothing else in the persona matters.
The validation signal — how they decide a solution is trustworthy. Some segments need peer reviews, others need a free trial, others want a direct sales conversation. Knowing this shapes the sales motion.
How to use the assumptions list
The prompt instructs the model to distinguish between claims grounded in the data you supplied and assumptions it has made to fill gaps. The assumptions list is the most valuable output. Take it into your next round of customer interviews and test each assumption directly. For example, if the model assumes “primary concern is time savings,” your next five customer calls should ask “what matters most to you about this?” and track whether time actually dominates. Replace assumptions with validated facts as you learn, and re-run the builder to get a sharper draft.
Tips and examples
Feed the model real signal wherever you can: even three sentences of support-ticket themes will dramatically sharpen the output versus an empty data field. Build one persona per distinct segment rather than a single averaged one — a price-sensitive solo founder and an enterprise procurement lead need different prompts. After running the prompt, copy the “assumptions to validate” list straight into your next round of customer interviews; that list is the most valuable output. Re-run the builder quarterly as you learn more, replacing assumptions with validated facts so each persona gets more accurate over time.