Deep Dive · XiaoHu Explains

The Top PM's AI Leverage Ladder: Two Paths to Boost PM Productivity, Plus 3 Ready-to-Use Prompts

An instructor who has trained 30,000+ PMs breaks down two AI leverage ladders — from copy-paste to end-to-end delivery, from web prototypes to production PRs.

One-Minute Takeaway
  • Colin Matthews breaks down how product managers use AI into three types of leverage — personal, product, and systems — each with three rungs. Climb a rung, and AI does more of the work while you review less.
  • Personal leverage: Rung 1, AI only drafts for you and you copy-paste manually; Rung 2, AI produces artifacts directly, like financial models; Rung 3, AI connects to real tools like PostHog and runs an entire task end to end.
  • Product leverage: Rung 1, use web prototyping tools to quickly validate ideas, but the code is unusable; Rung 2, use Claude Code / Codex connected directly to the real codebase to prototype; Rung 3, let an agent submit a mergeable PR straight to production.
  • Every key rung comes with the author's actual, battle-tested prompts — for a cost-comparison model, a PostHog retention analysis, and building a prototype-only codebase — reproduced verbatim here, ready to copy.
  • The framework's third ladder, systems leverage, is only defined at the outset — the body only fully develops personal leverage and product leverage; systems leverage never gets its three-rung breakdown.
1 Framework Overview

PMs using AI actually climb three separate ladders

Lenny's Newsletter published a guest post on June 30, 2026. The author, Colin Matthews, is a longtime product leader and serial founder who has independently shipped 10+ SaaS products. The piece was written to launch a PM course he built with Lenny.

Colin Matthews maps how product managers use AI onto three ladders — personal leverage, product leverage, and systems leverage — each climbing from Rung 1 to Rung 3. The higher you go, the more AI completes for you, and the less hands-on review you need to do.

📌

Why it's worth reading: The author has trained over 30,000 product managers, with client teams from OpenAI, Google, Stripe, Figma, and Microsoft, and has independently shipped 10+ SaaS products himself. What he describes are usages he's actually battle-tested, with the exact original prompt attached at every rung.

30,000+
Number of PMs the author has trained; clients include OpenAI, Google, Stripe, Figma, Microsoft
10+
Number of SaaS products the author has independently shipped
2 × 3
Ladders fully broken down in the original: personal leverage + product leverage, 3 rungs each; systems leverage is only mentioned as a concept, never developed

Each of the three leverage types handles one job: personal leverage helps you get through your own to-do list; product leverage helps you ship the right thing faster; systems leverage turns your AI usage into repeatable steps that reliably produce high-quality results. The matrix below is the backbone of this piece — the columns are the three rungs, the rows are the three leverage types.

Rung 1
Assist
Rung 2
Delegate
Rung 3
Full Handoff
Personal
Leverage
Rung 1
Use AI to write text — PRDs, Jira tickets, emails — then copy-paste it yourself
Rung 2
Have AI produce finished artifacts directly: slide decks, Excel models, small prototypes
Rung 3 · Core
Connect it to real tools so it runs an entire task end to end and checks its own work
Product
Leverage
Rung 1
Validate ideas with web prototyping tools; the code has nothing to do with the real product
Rung 2 · Core
Prototype directly in the real codebase, using real components and real styles
Rung 3
Let the agent submit a mergeable PR straight to production
Systems
Leverage
Not covered
The original only defines the concept
Not covered
Turning usage into repeatable steps
Not covered
Consistently getting high-quality results
← You review a lot, AI does little　　further right means AI completes more for you and you review less　　AI does a lot, you review a little →

The author has one important caveat about this ladder: not everything needs to climb to the top rung. Which rung fits best depends on how deep the task at hand should actually use AI. Below is the original framework diagram from his article.

Diagram of the three leverage ladders: personal, product, and systems, each with three rungs
The author's three leverage ladders: personal / product / systems, each running from assist to end-to-end. Source: Colin Matthews / Lenny's Newsletter
What does "leverage ladder" mean

It's a ruler that measures how deep you're using AI for a given type of work. At Rung 1, AI assists while you do the work yourself; at Rung 2, you hand off the task and AI produces a result for you to review; at Rung 3, AI works through several steps to finish the whole thing and checks its own results. The higher you climb, the more time you free up.

2 Personal Leverage · Rungs 1 & 2

From drafting for you to handing you a finished artifact