AI PM Maturity Assessment

A skill for benchmarking how effectively a product management team uses AI across the full product lifecycle. Works with Claude Code and GitHub Copilot.

What It Does

Produces an honest, evidence-based assessment of a PM team’s AI maturity. Scores 12 PM lifecycle phases on a 0–5 scale, identifies strengths and gaps, and generates concrete next steps — grounded in real workspace artifacts, not self-reported estimates.

The 12 Lifecycle Phases

1. Product Discovery & Market Research
2. Analyzing Customer Feedback & User Research
3. Competitive Analysis & Market Intelligence
4. Requirements Gathering & PRD Writing
5. Metrics Definition & Success Criteria
6. Design & Prototyping
7. Stakeholder Alignment & Communication
8. Developer Communication & Technical Specs
9. Building, Coding & Implementation
10. Testing, QA & AI Evals
11. Launch, Growth & Distribution
12. Team Structure, Hiring & AI Adoption

Scoring Scale

Installation

Claude Code

Copy the skill into your project’s .claude/skills/ directory:

# From your project root
git clone https://github.com/fitzgeraldsteele/ai-pm-maturity-assessment-skill /tmp/ai-pm-skill
cp -r /tmp/ai-pm-skill/.claude/skills/ai-pm-maturity-assessment .claude/skills/

The skill is then available in any Claude Code session within that project.

GitHub Copilot (VS Code)

Copilot recognizes skills in .claude/skills/ automatically — same install as above. The skill will be available as a slash command in Copilot Chat once the files are present in your repo.

Alternatively, copy the agent definition for a Copilot-optimized experience:

cp /tmp/ai-pm-skill/.github/agents/ai-pm-assessor.agent.md .github/agents/

Usage

Once installed, invoke it when you want to:

• Benchmark a PM team’s AI adoption
• Identify the highest-leverage gaps in AI-assisted workflows
• Plan where to invest in AI tooling for a product team
• Compare practices against industry benchmarks from leading AI-native companies

Files

• .claude/skills/ai-pm-maturity-assessment/SKILL.md — Skill definition, rubric, assessment workflow, and output format (works with Claude Code and GitHub Copilot)
• .github/agents/ai-pm-assessor.agent.md — Custom agent definition for GitHub Copilot
• AI in Product Management — Lifecycle Rubric & Benchmark Guide.md — Full rubric with sourcing and best-practice signals for each phase

Sources

The rubric synthesizes practices from 40+ episodes of Lenny’s Podcast (Jan 2025–Apr 2026), drawing from practitioners at Anthropic, OpenAI, Microsoft, LinkedIn, Airtable, Vercel, Reforge, Figma, Intercom, and others.