Boards are fielding more AI questions than ever. In the past year alone, the volume of AI-related agenda items has surged across industries — and that trend shows no sign of slowing.
Most of the conversations I see stay stuck at the technical layer: model types, data privacy architecture, vendor security certifications. These aren’t the wrong things to discuss. But too often they become a way to feel rigorous without being accountable. Boards walk away feeling briefed when they should be walking away with answers.
The governance question I’d push every board to answer first is much simpler than any of that: what decisions are being influenced by AI output, and who catches it when the output is wrong?
Not who owns the AI strategy. Not which committee reviews the vendor contracts. Not who approved the implementation roadmap. Who catches the errors — and how fast?
That’s the question that reveals whether oversight is real or decorative.
If the board can’t answer it for even one material workflow — a credit decision, hiring screen, a customer escalation, a compliance flag — the oversight gap is already open. And it’s open right now, not in some future quarter when the governance framework is finalized. The technical briefings can come later. This question can’t.
There’s nothing novel about this principle. Good governance has always started with accountability before complexity. You identify who is responsible when something goes wrong before you optimize for when things go right. Audit committees didn’t emerge because accounting got interesting — they emerged because the cost of errors without accountability was too high.
AI doesn’t change that principle. It just raises the stakes. When boards skip accountability and head straight for architecture, they’re not avoiding a hard conversation. They’re deferring a more expensive one.
