Sachin Rekhi: how I use AI as a product manager

Sachin Rekhi spent over a decade as a product manager at LinkedIn before founding Notejoy, a notes platform for teams. In May 2026 he published a detailed guide on his Substack walking through ten core PM deliverables and assessing, for each one, how much AI actually changes the work.

The framework covers four dimensions: vision, strategy, design, and execution. Within each, specific deliverables are evaluated on whether AI genuinely speeds things up, produces outputs that require heavy editing, or adds little that a PM couldn’t do alone. Knowing where AI helps and where it doesn’t changes how a team allocates time.

Prototyping and research: clearest gains

Rekhi identifies prototyping as the area most changed by AI. Tools like Cursor and v0 can turn a written description into an interactive, clickable prototype in a single session — work that previously depended on a designer’s time and multiple review cycles. For customer insights, AI works well at synthesizing feedback from surveys, support tickets, or interviews into thematic patterns: not perfectly, but accurately enough to replace large parts of manual tagging. Data analysis shows similar returns, as conversational queries can now produce SQL and visualizations without pulling in a data analyst. Stakeholder communication tasks — meeting notes, status updates, executive summaries, slide decks — are also strong candidates for AI-assisted drafting.

Strategy and vision: where human judgment holds

The piece is equally honest about where AI produces convincing but shallow output. Product vision ideation is the clearest case. Generating an ambitious direction requires what Rekhi calls “taste” — an internalized sense of what good looks like, built over years of working on real products. AI-generated vision statements tend to be plausible-sounding generalizations, not the kind of sharp, non-obvious positioning that actually shapes a team’s decisions. Strategy has the same problem: one-shot AI strategies lack the contrarian perspective that makes a direction defensible against market alternatives. Roadmap prioritization is a third area where AI counts and categorizes but cannot weigh the nuanced tradeoffs that reflect what a product actually stands for.

AI as a thinking partner, not a decision-maker

The most useful concept in the piece is the idea that AI amplifies existing judgment rather than replacing it. Rekhi describes seeding conversations with Claude and ChatGPT with specific constraints and context — for instance, when working through a pricing challenge — and using the outputs as raw material to react to, not as answers to accept. The distinction between using AI to explore options and using it to execute decisions keeps the quality of strategic work in human hands.

Who this is for

Senior product managers who are building AI habits and want to prioritize where to focus will find the framework directly applicable. Team leads deciding which workflows to hand to AI — and which to protect from it — will also find it practical.