Five questions for the board’s AI agenda

1. Which assumptions in our strategy become weaker as AI improves?

An AI briefing should help the board examine its strategy. Where could customers start expecting a different service, or competitors operate at a different cost? Management should explain which assumptions it’s watching and how a change would affect its choices.

2. What needs to change before we see the promised value?

Faster document review may do little for customer onboarding if every case still waits for the same approval. Ask which changes to the workflow the expected benefit depends on, and who can authorise them. If it’s still an experiment, agree what needs to be learned before investing more.

3. What can AI decide or do, and how would we know if it went beyond that?

“Human oversight” leaves a lot unexplained. Which actions need approval? How would a failure be detected, and who could intervene? The answers should fit the authority given to the system. Reviewing a tool’s recommendation is different from overseeing an agent that can act on it.

4. How will we change the organisation so this can work?

Management should explain how the proposed change is reflected in budgets, responsibilities and incentives. If AI frees up capacity, where will that capacity go? Where the business case depends on people working differently, someone needs to own the transition from how they work today.

5. What should we change in our own board work?

AI may help us prepare for meetings, compare scenarios or question assumptions. We need to agree how we’ll use it with confidential material and check its analysis. We also need enough understanding to challenge management’s answers, including knowing where we need to learn more.

These questions build on the five boardroom AI topics identified by KPMG’s Per Edin, as summarised in Enterprise AI Executive’s briefing.