The AI-Native PM: Compress Months of Work into Days
Scher Rafay
Staff PM @ Meta | Wharton | Builder
Be the PM who ships fast and doesn't produce slop.
You wrote the spec four weeks ago. Your engineer "kind of" built it. Your skip-level wants a quick sync. It's now Tuesday and you're still six alignment meetings away from shipping.
You've tried the obvious moves to use AI. But the work takes the same time. Becoming an AI-Native PM, and actually seeing the benefits, requires you to rethink the process.
This workshop teaches the PM job differently:
Write a PRD in 10 minutes, sometimes after you've already built the prototype
Synthesize 30 user interviews in an afternoon
Generate 20 UX variations of an idea, and pick one with conviction
Design evals without an ML team holding your hand
Simulate your exec team's objections before the review, not in the room
Build prototypes when writing a spec is slower than building
Avoid AI tells in your work so it reads as yours
Generate fast with AI, judge well, ship work that's good. Speed is the upgrade. Taste is still your job.
Bigger payoff: shorter days, agents handling what you hate, more time on the work that drew you to product management.
I built this workflow at Meta and taught dozens of PM workshops internally. This isn't another "become an AI PM" course. This teaches you the job, better, with AI.
What you’ll learn
Move from spec-and-wait to ship-and-iterate, with AI doing the heavy lifting and your taste running the show.
How I built this workflow at 1.7B DAU scale and what the broader org actually adopted
A diagnostic framework for spotting which PM tasks are AI-ready vs. judgment-only vs. don't-touch-it
Your own AI-native PM operating system, sketched during the workshop and yours to keep
Live demo: I write a real PRD in 10 minutes using Claude Code + NotebookLM, prompts shared
The inversion: when writing the PRD after the prototype is faster, and what changes in your workflow
A PRD scaffold you can re-use immediately. No more starting from a blank doc
Synthesize 30 user interviews into themes and next steps in a single session using Claude + NotebookLM
Design evals for AI features without an ML team holding your hand
Generate 20 UX variations and develop conviction about the right one in 30 minutes
Steel-man your skeptical exec's objections before the review using Claude as the devil's advocate
Synthesize XFN partner POVs into a decision doc that addresses each stakeholder's actual concern
Draft tone-perfect responses to high-stakes emails/threads (the ones you stress about all evening)
Live build: I prototype a real AI feature in Claude Code in 60 minutes, no engineer needed
The decision framework: when to spec, when to prototype, when to just send the working demo
Your own working prototype, shipped during the workshop and yours to demo Monday morning
The AI tells reviewers learn to spot (em-dashes, "delve," buzzword rhythms) and how to avoid them
A voice-matching prompt pattern that makes Claude write like you, not like Claude
The 60-second taste pass that catches AI slop before anyone else sees your work
Workshop agenda
The AI-native PM thesis + case study
I'll show you the AI-native workflow I built, the one I've taught internally. Sets the frame for the rest of the day.
AI ideation without slop: generate fast, judge well
Use AI to generate 20 options fast, then apply your judgment to cut to the 3 that matter. The framework that protects your taste.
AI-native PRDs: from 4 hours to 10 minutes
Live demo of writing a real PRD with Claude Code + NotebookLM. Plus the inversion most PMs miss: when writing the PRD after the prototype is actually faster. Lab + template included.
AI-native research synthesis + eval design
Two demos back-to-back. Synthesizing 30 user interviews into themes and a roadmap in an afternoon with Claude. Then designing evals for AI features without an ML team hand-holding. Both with labs.
Lunch Break
30-minute break to recharge. Come back ready for the afternoon politics block.
AI for PM politics: surviving exec reviews and XFN partner pushback
Steel-man your skeptical exec's objections before the meeting using Claude as devil's advocate. Synthesize XFN partner POVs into decision docs that actually land.
Build your own prototype, live
I prototype a real AI feature in Claude Code in front of you in 60 minutes. Then you build one with help. You leave with a working prototype to demo Monday morning.
Final shares + Q&A
Everyone shows what they built. Open Q&A on anything from the day: workflow, tools, politics, prototype debugging. Recording shared after for lifetime reference.
Learn directly from Scher
Scher Rafay
Staff PM @ Meta | Ship AI to 1.7B DAU | Wharton
Who this workshop is for
PMs who feel slow watching indie builders ship with AI on weekends, and want to compress cycles without losing their taste.
PMs at startups whose teams are going AI-native, and need to model what good AI-native PM work looks like.
PMs who don't build themselves anymore, but need to spot the difference between good AI-native work and slop.
What's included
Live sessions
Learn directly from Scher Rafay in a real-time, interactive format.
Lifetime access
Go back to course content and recordings whenever you need to.
Community of peers
Stay accountable and share insights with like-minded professionals.
Certificate of completion
Share your new skills with your employer or on LinkedIn.
Maven Guarantee
Your purchase is backed by the Maven Guarantee.
Frequently asked questions
Maven for Teams
Reimbursement
Get your company to pay
Everything L&D needs: email template, receipts, and certificate of completion.
Get reimbursedTeam discount
Learn with your teammates
Save 20%+ when 2 or more teammates enroll in the same cohort.
Save 20%+ with a teamPrivate cohort
Run a cohort for your org
A dedicated cohort with a custom schedule and curriculum, tailored to your team.
Book a private cohort$499
USD