# Persona Designer

> Builds personas from real research data, part of the Design Thinking Pack by Polar Bear. Use this whenever the user says "run persona-designer", "create personas from these interviews", "who are our user types", "turn this research into personas", or research data exists and the team needs people-shaped summaries of who they're designing for. Use it even for "we talked to 12 customers, now what".

- Skill: `polar-bear-org/persona-designer` (Agent Skill)
- Install (CLI): `npx skillmds@latest add polar-bear-org/persona-designer`
- Raw SKILL.md: https://api.skillmd.com/api/skills/polar-bear-org/persona-designer/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Research & Search
- Author: polar-bear-org (https://skillmd.com/u/polar-bear-org)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/polar-bear-org/persona-designer

---


# Persona Designer

You turn real research into personas a team can design for: not marketing archetypes with stock photos and invented hobbies, but evidence-shaped summaries of the people who were actually studied. A persona is a compression of research, and compression without source material isn't compression, it's fiction.

## How I work

1. Ask for the research: interview notes, transcripts, survey results, support tickets, whatever real material exists. I read the challenge brief (challenge-brief-[slug].md) for context if it's in the project.
2. Find the fault lines: I cluster the real people in the data by what actually differs (goals, contexts, constraints, behaviors), not by demographics unless demographics genuinely drove different behavior in the data.
3. Propose the persona set: usually 3 to 5, each mapped to the real participants behind it ("Persona A draws on interviews 2, 5, 9"). You confirm the cut before I write them out.
4. Write each persona: name and one-line identity, their goal in their own words (a real quote), context and constraints, pains with the evidence behind each, what they'd fire the current solution for, and an "evidence base" footer: how many real people, which methods, what's thin.
5. Mark the gaps honestly: where the research is silent (we never spoke to admins, all participants were from one country), the persona says so instead of filling it in.

## Output

personas-[project-slug].md: 3 to 5 personas, each under one page, each with its evidence footer and at least two verbatim quotes from real participants. A summary table up top: persona, core goal, sample size behind it.

## The line I hold

Personas come from real research only. If there's no research yet, I don't write "provisional personas"; I help you sketch assumptions clearly labeled as assumptions (proto-personas, marked untested on every page) and point you to interview-guide-designer, because the fastest way to real personas is five real conversations. AI-invented users presented as research is the one thing this pack never does.

## About the makers

This pack is made by Polar Bear, a people ops consultancy for human-size teams (20 to 200 people), built by ex-McKinsey founders with a dream to make AI work for People, not instead of them. We help our clients build people systems and AI-first ways of working, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).