Product Growth: How to build product strategy in the age of AI — step by step

Aakash Gupta delivered this keynote at Northeastern University in March 2026 and made it available publicly through his Product Growth channel. The central premise: when AI coding tools like Claude Code and Cursor allow engineers to ship in hours what used to take days, the bottleneck in product development stops being execution and becomes direction. Teams that can move fast but lack a clear strategy are not more productive — they are just wrong more quickly.

Who it is for

This talk is most relevant for PMs and product leaders at companies where AI-assisted development is already part of the engineering workflow, or where that transition is underway. It is also useful for anyone in a product role who has noticed that their team’s ability to execute has outpaced their ability to agree on what to build. No particular technical background is required.

Key takeaways

1. Strategy clarity is the new execution bottleneck. Gupta opens with a concrete observation: when asked about their product strategy, nine out of ten engineers on a given team cannot explain it in 30 seconds. This gap becomes more costly as execution speed increases, because misaligned fast teams produce misaligned fast output. The problem is not intelligence or effort — it is that strategy rarely gets communicated in a form people can actually act on.

2. The 7-step strategy framework: Objective → Users → Superpowers → Vision → Pillars → Impact → Roadmap. Each step builds on the previous, and the sequence matters. Building the roadmap before establishing a clear vision produces a list of features with no strategic logic connecting them. Gupta walks through each step with examples and the specific questions each step is meant to answer.

3. Common strategy failures are structural, not analytical. Gupta identifies four recurring failure modes: strategies that are too long to remember, too vague to act on, disconnected from daily work, and static rather than updated as conditions change. These are communication and organizational problems, not problems that a better framework alone will fix.

4. AI is a useful tool for stress-testing strategy, not generating it. Gupta shows how to use AI models to surface counterarguments to a proposed strategy, synthesize research quickly, and identify gaps in reasoning. He is direct that original strategic thinking needs to come from people who understand the organization, the users, and the competitive context.

5. Fast execution raises the stakes for strategic judgment. The keynote’s closing argument is that AI-assisted teams need more rigorous strategy work, not less. When going in the wrong direction takes hours to produce output rather than weeks, the cost of being wrong accumulates faster, and course corrections become more disruptive.

Worth watching if

You lead a product team where engineering velocity has increased substantially due to AI coding tools, and you have noticed that clarity about direction has not kept pace — or you want a practical framework to run a strategy-setting exercise with your team.