Working notes on AI-native product management.
How to use AI agents — Claude Code, Codex, Antigravity — for real PM work: PRDs, discovery, strategy and exec communication. Written by a PM who ships with these tools every week.
ChatGPT vs Claude Code for PM Work: An Honest Comparison
They're not competing for the same job. An honest, no-hype comparison of ChatGPT and Claude Code for product management work, and how to decide which one your week needs.
Guide · EvalsHow PMs Should Write Evals for LLM Features (v0 Guide)
Most AI features ship on vibes, not evals. A lean v0 guide to writing eval plans and running error analysis on real traces — a copy-paste prompt included.
Guide · QualityFrom PRD to QA Test Scenarios with AI
QA usually reverse-engineers what a PRD meant. How to generate test scenarios from a PRD with an AI agent, traced to specific requirements, edge cases included — a copy-paste prompt.
Guide · DeliveryEvent Tracking Specs with AI: From User Story to Taxonomy
Analytics usually gets bolted on after launch, inconsistently. How to spec event tracking with an AI agent at the same time as the feature — a copy-paste prompt included.
Guide · Exec CommunicationDecision Briefs That Survive the Exec Room: AI Pushback Simulation
Most decision briefs fail live, not on the page. How to pressure-test a brief with an AI pushback simulation before you present it — a copy-paste prompt included.
Guide · DeliveryHow to Write Jira Epics with AI (with Cagan's Four Risks)
Most Jira epics are just repackaged feature lists. How to write epics with an AI agent structured around Cagan's four risks and sequenced by what could kill the bet — a copy-paste prompt included.
Guide · StrategyProduct Strategy Sparring with AI: Beyond the Blank Template
Why a strategy template doesn't produce a strategy, and how to use an AI agent as a sparring partner that diagnoses, pushes back, and forces real trade-offs — a copy-paste prompt included.
Listicle · The command library16 Claude Code Commands Every PM Should Steal
A working list of the 16 commands that cover a PM's actual week, grouped by discovery, strategy, delivery, quality, evals and exec communication — steal what you need this week.
Guide · DiscoveryBuilding Opportunity Solution Trees with AI (Teresa Torres Method)
How to build an Opportunity Solution Tree with an AI agent grounded in real interview evidence instead of workshop guesses — a copy-paste prompt and how to prune it without losing the trail.
Guide · DiscoveryHow to Analyze User Interviews with AI Without Losing the Insight
Why "summarize these interviews" flattens your best signal, how to ground an agent in your hypotheses first, and a copy-paste prompt that preserves outliers instead of averaging them away.
Guide · Working BackwardsWriting an Amazon-Style PR/FAQ with AI (Working Backwards)
How to write a Working Backwards press release with an AI agent: the PR/FAQ structure, a copy-paste prompt, and how to stress-test the FAQ before you build.
Guide · FoundationsWhy AI Prompt Packs Fail PMs (and What Works Instead)
Most "AI prompts for product managers" lists get used twice and abandoned. Why prompt packs hit a ceiling fast, and how a context system fixes it — with a side-by-side prompt example.
Guide · FoundationsClaude Code for Product Managers: The Complete Non-Engineer's Guide
Installing Claude Code with zero coding background, setting up context in fifteen minutes, and running your first PRD session — a step-by-step guide for PMs.
Guide · DeliveryHow to Write User Stories with AI (Jira-Ready, with Acceptance Criteria)
INVEST-format stories, testable Given/When/Then acceptance criteria, and a copy-paste prompt — how to generate Jira-ready stories an engineering team can actually build from.
Guide · DeliveryHow to Write a PRD with Claude Code (a PM's Step-by-Step Guide)
Why coding agents beat chatbots for PRDs, how to ground the agent in your product context, and a copy-paste starting prompt — no engineering skills required.
Guide · FoundationsAgentic Product Management: A Practical Guide for PMs
What "agentic PM" actually means beyond the buzzword, why context beats prompts, and a concrete five-step path from chatbot user to agent-native product manager.