20 Nov · Simon Willison · Product Strategy
Be the PM who ships AI. Learn, practise and prove the skills, and grow your career.
Are you AI PM ready?
AI products fail in ways ordinary products don’t. Can you spot the failure before your users do?
Five real scenarios · about 2 minutes · no sign-up
Readiness check
Question 1 of 5Evals and metrics
You ship a new model and prompt. Your offline LLM-judge score rises from 0.82 to 0.89. The same week, the share of users who edit or throw away the output rises 4 points.
What do you do?
Tap an answer. Nothing to type.
Close the gaps, skill by skill
Each skill is a few short pages, with the PM call at every step.
Explain how LLMs and agents work
Talk to engineers in their terms.
AI fundamentals · AI agents · Multi-agent systems · AI tools & prompting
Spot and scope an AI feature
Decide where AI earns its place, then cut it to ship.
AI opportunity spotting · Scoping AI products · Discovery & requirements
Evaluate an LLM feature
Build the eval set and know when quality is good enough.
Ship AI safely
Find the failure modes before users do.
Run an AI product after launch
Roadmap under uncertainty; manage drift and cost.
Managing AI products · AI roadmapping · Process optimisation
Master the PM fundamentals
The frameworks every PM interview still assumes.
Agile & Scrum · Prioritisation · Estimation · ADO & user stories · Product & team KPIs · PM documents
Stuck on a term? The AI glossary has it.
Stay current
One AI development at a time, explained for product managers, with what to do about it.
Open the AI PM DigestProve it, then make your move
Practise like the real interview
Your career
Know what the role owns, then turn your background into a story about one AI decision you can defend.