← ClaudeAtlas

user-persona-builderlisted

Build evidence-based user personas and Jobs-to-be-Done (JTBD) statements from interview notes, survey results, reviews, CRM or sales notes, and support tickets. Use when a PM, PMM, UX designer or founder needs personas, an ideal customer profile (ICP), JTBD, or the "target users" section of a PRD or marketing plan. Every persona trait is traced to evidence or labeled as an assumption.
lingegrace/claude-research-skills · ★ 0 · Web & Frontend · score 72
Install: claude install-skill lingegrace/claude-research-skills
# User Persona Builder Input: whatever the team already knows about its users (interview notes, survey data, reviews, CRM fields, sales call notes). Output: 2–4 personas with jobs, pains, triggers and buying criteria, each one traceable to real evidence, so they aren't invented demographics. ## How to work 1. **Clarify in one message.** Ask: - **Goal:** what are the personas for? (a PRD, messaging, sales targeting, UX design) - **Product and market:** what the product does, B2B or B2C, the market and region - **Evidence on hand:** which materials they can share, and roughly how many interviews or responses - **Known segments:** any segments the team already uses (by role, company size, plan) Done when you know the goal and have at least one piece of real evidence. With no evidence at all, say the result will be a **hypothesis persona** to validate, and label it that way. 2. **Extract signals.** From each source, pull out: role and context, the **job** they're trying to get done, the **trigger** that made them look for a solution, **pains** and workarounds, **desired outcomes**, **buying criteria** and who else decides, and **objections**. Keep the source of each signal. Done when every signal has a source. 3. **Cluster.** Group people by job and situation, not by age or gender. Demographics belong in a persona only if they change behavior. Done when each cluster has a clear, different job, and you have 2–4 clusters. 4. **Write the personas.** One card pe