The Difference Between an AI Strategy and an AI Plan

Most companies think they have an AI strategy. What they have is a project list with "AI" written at the top.

That's not a small distinction.

A strategy answers why and where. A plan answers how and when. Almost every company I talk to has built a plan (sometimes a very detailed one) and is calling it a strategy. The two are not interchangeable, and confusing them is how AI initiatives stall, get defunded, or quietly get absorbed back into IT roadmaps.

What a plan looks like

A plan tells you what to build and in what order. Milestones, owners, a Gantt chart if you're unlucky.

None of that is wrong. You need a plan. But a plan without a strategy underneath it is organized activity. You're executing without a compass.

I've watched companies run their AI pilot, hit a wall, and then look at their plan to figure out what comes next. The plan says "Phase 2: Expand to three additional departments." But expand what: a pilot that didn't work? And why those three departments — because they were next on the list, or because they're the ones where this approach has a real shot?

A plan can't answer those questions. That's what strategy is for.

What a strategy looks like

A strategy starts with a position: where you're trying to win, and where you're willing not to compete. What do you believe about how AI creates value in your industry that competitors aren't acting on yet?

That bet informs every subsequent decision: which use cases to prioritize and where to develop internal capability versus buying it. Strategy is what lets you say no; a plan only tells you what to say yes to.

For most companies, this means having an honest conversation about what they're actually trying to become. Not "we want to use AI to be more efficient." That's everyone. Efficiency is a plan, not a position. The companies with real AI traction have made a specific bet: AI is going to shift value creation toward X, and we're going to be early in X.

Why AI makes this more important, not less

Software roadmaps can survive without a clear strategy underneath them. Ship features, watch usage, course-correct. The feedback loop is tight.

AI is different. You're making bets on models, datasets, and team capabilities that take time to compound. A bad use case choice doesn't waste a sprint — it can set you back six months and create organizational skepticism that's hard to undo.

Enterprise AI adoption is especially fragile in the early stages: the first failed initiative becomes the proof point skeptics cite for the next two years. "We tried AI and it didn't work" is one of the most expensive sentences inside any large company. Strategy is your protection against that. When the why and the where are clear, a failed pilot tells you something useful. You learn and adjust.

Strategy tells you what to learn from a setback; a plan just tells you what you were supposed to do.

The test

Here's how I tell whether a company has a strategy or a plan: I ask what they'd stop doing in AI if they had to cut 50% of the budget tomorrow.

If they can answer quickly, they have a strategy — they know which bets matter and why.

If it becomes a committee discussion about political sensitivity, they have a plan: a set of commitments with no real hierarchy of importance.

Getting from a plan to a strategy isn't complicated. You need honest answers to a couple of questions: Where are we trying to win? Why does AI specifically change the game there? Write that down somewhere people can actually read it. Then build your plan underneath it.

A plan without a strategy is just a calendar.