What about AI?

Faster cars don't make better drivers.

There's a popular narrative right now that speed is the only moat in an AI-first era. That's backward. Speed is a requirement, not an advantage. An easy way to check is to use the opposite test: is the opposite of "move faster with AI tools" clearly a bad idea? Yes. Assuming we're talking about building technology, no sane competitor would choose NOT to use AI. Using AI is operational table stakes, not a differentiating choice. Everyone should and will use AI, which means it's neither a strategy nor a moat.

Michael Porter nailed this in 1996, writing: "caught up in the race for operational effectiveness, many managers simply do not understand the need to have a strategy." Best practices spread rapidly and everyone gets better. No one gains sustainable advantage. This is exactly what's happening with "speed is the only moat."

As Brian Balfour puts it, "think of speed like the first stage of a rocket. It's absolutely essential for getting off the ground, but useless if you don't have subsequent stages to reach orbit." AI-driven execution speed is a rocket. Strategy is a navigation system. You need both.

So, what does AI actually change for product management? It shifts the bottleneck: when everyone can build things fast, the constraint moves from "can we ship it?" to "do we know what to build, and why?" That's a strategy problem, not a delivery problem, and it increases the cost of going in the wrong direction. Without clear strategic thinking, AI just gives you accelerated chaos. You go the wrong way, faster.

Here's what shifts in practice:

• You can run faster cycles. If you can build and ship faster, you can learn faster. Consider trying cycles shorter than a quarter. But don't confuse shipping speed with learning speed, which is the real bottleneck. You still need time to observe results and figure out what to do based on actuals.
• Context moves faster. When everyone can build faster, markets change more quickly. You have to stay on top of the shifting context. An evolving strategy becomes more important, not less.
• Having a clear Value Thesis is even more important, especially defensibility. When speed is the norm, the choices you make are the only remaining multiplier on your results.

Where AI helps your strategy work

It's helpful to think about AI in two ways: how it shows up in your actual product, and in your process.

The role of AI within your product will constantly evolve as tools and models improve. What's more useful, and durable, is how AI helps in the process of your work as a strategist. It can play five distinct roles: (1) a researcher that gathers and enriches data, (2) an interpreter that turns data into insights, (3) a thought partner that brainstorms and checks your bias, (4) a simulator that models scenarios, and (5) a communicator that improves your drafts.

What AI does throughout the work is take on the laborious parts so you can spend your energy where it matters: focus, creativity, judgment, sense-making, and bringing people along for the ride. It's most useful in the Craft phase (Part 1) and tapers off in Deploy and Manage (Part 2), which are primarily human-to-human work.

AI can't do your hard thinking for you. The choices remain yours. Don't just build a faster car. Become a better driver.