🎯 AI Customer Discovery Skills

Turn raw customer signal into validated product opportunities — in minutes, not weeks

Maintained by Varun Kulkarni

---

🔭 The Discovery Flywheel

flowchart LR
    subgraph SIGNAL["📥 RAW SIGNAL"]
        S1["Support tickets"]
        S2["Sales call notes"]
        S3["NPS / surveys"]
        S4["Interviews"]
    end

    subgraph SKILLS["🧠 DISCOVERY SKILLS"]
        K1["feedback-prioritizer<br/>RSCF ranking"]
        K2["competitive-analyzer<br/>buyer-weighted teardown"]
        K3["assumption-mapper<br/>K/B/H × Crit/Low"]
        K4["north-star-metric-finder<br/>5-criteria pick"]
        K5["jtbd-extractor<br/>functional / emotional / social jobs"]
    end

    subgraph OUTPUT["🎯 VALIDATED OPPORTUNITY"]
        O1["Ranked backlog"]
        O2["Defensible competitive POV"]
        O3["Ranked test plan"]
        O4["North Star + input metrics"]
        O5["Scored JTBD opportunities"]
    end

    SIGNAL --> K1 --> O1
    SIGNAL --> K2 --> O2
    SIGNAL --> K3 --> O3
    SIGNAL --> K4 --> O4
    SIGNAL --> K5 --> O5

    classDef skill fill:#1a73e8,color:#fff,stroke:#1558b0,stroke-width:2px,rx:6,ry:6
    classDef signal fill:#fef7e0,color:#202124,stroke:#fbbc04,stroke-width:1px,rx:6,ry:6
    classDef output fill:#e6f4ea,color:#0d652d,stroke:#0d652d,stroke-width:2px,rx:6,ry:6
    class K1,K2,K3,K4,K5 skill
    class S1,S2,S3,S4 signal
    class O1,O2,O3,O4,O5 output

Loading

---

Why This Library Exists

Customer discovery is the first place product work goes wrong: signals get cherry-picked, opportunities get sized by gut feel, and personas get overfitted to whoever shouted loudest in the last interview. This library captures the structured workflows that turn raw signal into evidence — each skill is a self-contained markdown file that any compatible AI agent can load on demand.

---

⚡ Quickstart

# 1. Clone the repo
git clone https://github.com/varunk130/ai-customer-discovery-skills.git

# 2. Install skills globally for Claude Code
mkdir -p ~/.claude/skills
cp -r ai-customer-discovery-skills/skills/* ~/.claude/skills/

# 3. Restart Claude Code, then run a skill:
#      /feedback-prioritizer        — triage a backlog of support tickets
#      /competitive-analyzer        — score competitors on buyer dimensions
#      /assumption-mapper           — surface and rank bet-killing assumptions
#      /north-star-metric-finder    — pick a 2-year-horizon north star
#      /jtbd-extractor              — turn interviews into ranked JTBD statements

GitHub Copilot users: copy the same skills/ directory into .github/skills/ in any repo and invoke via natural language.

Project-local install: drop the skills into .claude/skills/ inside your project to scope them to one codebase.

---

📋 Skills Catalog

Skill	What it does	Use when
feedback-prioritizer	Triages raw customer feedback into a ranked list using the RSCF model, with an explicit Do Not Act list for vocal-minority signals	A backlog of tickets / interviews / NPS / sales notes is piling up and the team needs focus
competitive-analyzer	Runs a disciplined competitive teardown - picks 4-6 buyer-weighted dimensions, scores every competitor, and surfaces gap + risk maps	You need a defensible competitive analysis that changes a decision, not a 40-row feature grid
assumption-mapper	Surfaces hidden assumptions, classifies them Known / Believed / Hoped × Critical / High / Medium / Low, and outputs a ranked test plan	You're about to commit real investment to a bet and need to know what could kill it first
north-star-metric-finder	Identifies a candidate North Star Metric using five strict criteria, then maps the input metrics that drive it	You're picking the single metric that will steer two years of roadmap decisions
jtbd-extractor	Turns raw research (interviews, support, sales notes) into ranked Jobs-to-be-Done statements with functional / emotional / social jobs and opportunity scoring. Ships with a Python CLI + HTML renderer	You have a corpus of customer conversations and need structured, prioritized JTBD output you can actually act on

---

🗺 Roadmap

5 of 12 skills shipped. Additional skills planned: persona-validator, opportunity-sizer, switch-cost-analyzer, willingness-to-pay-tester, and more. Watch this repo for releases.

---

Related Work

Part of a portfolio of AI agent and skill libraries for product, GTM, and decision-making teams.

Discovery & research

• jtbd-extractor - Extract Jobs-to-be-Done statements from research, with opportunity scoring

Strategy & decisions

• claude-code-skills - 29 production-grade skills for finance, product, strategy, and game theory
• AI-Builder-Decision-Analyst - 11 skills that catch bad bets before you ship across DECIDE / BUILD / COMMUNICATE / LEARN

Go-to-market

• ai-gtm-skill-library - 31 opinionated GTM skills across the full discover -> renew lifecycle
• ai-marketing-claude-skills - 12 marketing-ops skills with scoring algorithms and statistical frameworks
• ai-partner-ecosystem-analysis - Deep research on any ISV, partner, or competitor with a 1-slide PPTX output

UX & design

• ai-ux-skill-library - 12 frameworks for designing UX for AI products, agents, and AI-powered experiences

Multi-agent demos

• ai-pm-agents-suite - 6-agent pipeline plus 3 standalone PM agents (decision engine, financial analyst, stakeholder translator) that turn customer feedback into strategy, PRDs, and comms
• ai-legal-agents-skills-os - Agentic operating system for legal work: one master agent, nine specialist skills, MCP + MCP Apps

Evaluation & operations

• AI-Eval-Skills - 6 skills to plan, generate, run, interpret, and triage AI agent evaluations
• ai-workflow-playbooks - 21 playbooks + 10 skills + 4 guardians + 5 runbooks across the 7-stage delivery pipeline

---

License

MIT — use freely, attribution appreciated.