The Board Wants an AI Strategy. That's the Wrong Ask.

Every board wants an AI strategy. The phrase has become routine in quarterly reviews, offsite agendas, and analyst calls. Executives who two years ago asked about digital transformation are now asking about artificial intelligence, often in the same breath, with the same urgency.

The ask is understandable. But it's the wrong question.

The Problem With "AI Strategy"

Asking for an AI strategy puts the technology first. It implies that the task is to find applications for a capability you've already decided to acquire — a logic that produces roadmaps full of proofs of concept and very few things that actually ship.

Strategy is about where to compete and how to win. AI is a set of techniques. Techniques don't have strategies; businesses do.

The companies that have made the most durable use of machine learning over the past decade didn't start with "what should we do with AI?" They started with specific, expensive problems — fraud detection, demand forecasting, content recommendation — and asked whether a model could outperform their existing approach. AI was the answer, not the question.

The Right Reframe

The question that gets real work done is: which of our highest-value problems are currently bottlenecked by human judgment at scale?

That framing does two things. First, it anchors the conversation in the business, not the technology. Second, it surfaces the problems where AI has a structural advantage — not because it's smarter than your team, but because it can apply a consistent decision at a volume and speed that human judgment cannot.

Call center routing, document classification, procurement anomaly detection, customer lifetime value scoring — these aren't glamorous. But they compound. A model that routes 40% of support tickets more accurately doesn't make a great press release; it shows up in operating leverage two quarters later.

What to Do on Monday

Three things that move faster than another strategy workshop:

Build a use-case backlog. Pull your top ten operational processes that involve high-volume, repetitive human judgment. Rank by decision volume and error cost. The top three are your AI candidates — not because AI is interesting, but because the improvement math works.

Define a success metric before you build anything. Not "improve accuracy" — a specific, measurable outcome tied to a business number. If you can't name it before the project starts, you won't be able to defend the investment when the year-two pressure arrives.

Assign an owner who reports to a P&L. Capability without accountability drifts into the lab. The difference between a successful AI deployment and a successful proof of concept is almost always organizational, not technical.

The board's instinct isn't wrong. The investment case for AI in enterprise operations is real. But the strategy question that deserves an answer isn't "how do we become an AI company?" It's "what would we do differently if we could make a million decisions a day as well as our best analyst makes one?"