Free Lesson
AI PM in 30 Minutes: Become an AI-Native PM
30 min
Jun 30, 2026 6:00 PM
What you'll learn
How to spec a feature when the output isn't predictable
Writing requirements for systems that guess. What "done" means when the same input can give different answers.
The metrics that tell you an AI feature is actually working
Why thumbs-up isn't enough. The signals that catch quality drops before users churn.
Live teardown: why a "smart" feature is annoying people
A real AI feature users hate. We diagnose it as a product problem, not a model one.
Why this topic matters
Most PMs treat AI like a normal feature with a fancy backend — spec it, hand it off, ship it. But probabilistic products break that playbook. The hard calls aren't technical; they're product calls: what's good enough, what happens when the model is wrong, and whether the feature should exist at all. The PMs getting hired now own those calls instead of routing them to engineering.
You'll learn from
Aki Wijesundara, PhD
AI Founder | Educator | Google AI Accelerator Alum
Manu Jayawardana
AI Advisor | Co-Founder & CEO at Krybe | Co-Founder of Snapdrum
Previously Students from
