AI strategy isn't a tools decision. It's a board decision.

There's a quiet category error happening in a lot of organisations right now. AI shows up as a line item in the technology budget, gets handed to IT or a digital team, and is managed like any other software purchase. Compare vendors, run a pilot, negotiate a licence. And then leadership wonders why, two years and several tools later, nothing fundamental has changed.

The error is treating a strategic question as a procurement one. AI isn't a tool you buy. It's a capability that, done well, changes your cost structure, your competitive position and sometimes your business model. That makes it a board-level question, and boards that delegate it entirely tend to get incrementalism when they needed transformation.

The delegation trap

I understand why it happens. AI feels technical, so it gets routed to the technical function. But routing strategy through procurement guarantees a procurement-shaped outcome: tools that optimise existing processes rather than questioning whether those processes should exist at all.

If your AI conversation is mostly about which vendor to pick, you're having the wrong conversation. The vendor is a footnote. The strategy is the story.

The most valuable AI moves I've seen weren't "we bought a better tool." They were "we realised this entire function could work differently." That realisation never comes from a vendor comparison. It comes from leadership asking hard questions about where the business is actually going.

Three questions only the board can answer

There are decisions in an AI agenda that simply cannot be delegated, because only the board has the mandate and the vantage point to make them.

1. How much risk are we willing to take? AI introduces new categories of risk, around data, accuracy, bias, regulation and reputation. The board's risk appetite sets the boundaries everyone else operates within. Leave it unstated and teams will either freeze or overreach.
2. Where are we willing to be bold? You can't transform everywhere at once. Choosing which parts of the business get genuine investment and which simply get efficiency tweaks is a capital-allocation decision. That's the board's job.
3. What do we stand for as we adopt this? The principles governing how you use AI, transparency with customers, fairness in decisions, how you treat people whose work changes, are reputational commitments. They belong to the people accountable for the institution, not to a project team.

What good governance looks like

Board-level engagement doesn't mean directors writing prompts or reviewing model architectures. It means setting direction, appetite and principles, then holding the executive accountable for delivering within them. The board sets the guardrails; the organisation drives within them.

The board's real role

I work with boards precisely because this is where the leverage is. Get the direction, the risk appetite and the principles right at the top, and the rest of the organisation has a frame to work within. Get them wrong, or never set them, and you get a scatter of disconnected pilots, each technically fine, none adding up to anything.

AI is the most consequential strategic shift most established businesses will navigate this decade. Treating it as a purchasing decision is like treating entering a new market as a stationery order. The tools matter, eventually. But the strategy comes first, and the strategy belongs in the room where the future of the business is actually decided.