What Boards Should Be Asking About AI

Most board-level conversations about AI still revolve around the wrong questions. "What is our AI strategy?" and "How much are we spending?" produce reassuring presentations and very little operational clarity. The boards getting real signal are asking a different set of questions.

1. Which decisions in this company are still being made on intuition that should be made on evidence? This forces leadership to identify where intelligence will create asymmetric advantage, rather than treating AI as a horizontal initiative.

2. What is the cost of latency in our most important workflows? If a credit decision takes three days, a hiring decision takes six weeks, or a procurement decision takes a quarter, that latency has a dollar value. AI investments should be benchmarked against it.

3. Who in this organization owns the data that the models depend on, and are they resourced for the role they are now playing? Data ownership has quietly become one of the most consequential leadership roles in the company.

4. What is our policy when a model is wrong? Disciplined programs have answers. Experimentation does not.

5. How are we measuring whether our people are becoming more capable, or simply more dependent? Enablement is the difference between durable advantage and rented intelligence.

Boards are not expected to operate the technology. They are expected to ensure the organization is being honest about where it is, disciplined about where capital is going, and accountable for the systems it deploys. The questions above are the ones that surface those answers.