Boards have read the same headlines you have: record AI spending, a 95% pilot failure rate, and CEOs who cannot find the benefit. So when an AI rebuild proposal reaches the boardroom, or the CFO's office, or the family council, the questions are not hostile; they are predictable, and predictable questions reward preparation. Twelve come up again and again. Prepare the answers before the meeting, because the preparation is most of the credibility.
On the thesis
1. Why now? AI has been around for years. They are asking whether the urgency is real or fashionable. Answer with the 2026 paradox data: spend tripled, bolt-on pilots failed at roughly 95%, and workflow redesign is the proven differentiator, with a compounding penalty for each year of delay.
2. What exactly are we betting on? They want a thesis, not enthusiasm. One sentence: AI does the operating work, we redesign the workflows, we measure quarterly. Then show the one-page diagnostic map.
3. Why will we succeed where 95% fail? Name what is structurally different: diagnose first, foundation first, one workflow end to end, measurement from day zero, phased gates. If your plan lacks those five, this question is where the proposal fails, and it should.
On the numbers
4. What does it cost, all in? Present a phased budget with the evidence required at each gate, and name the internal time cost, not just the external spend. An honest total buys more trust than a small one.
5. How do we measure return? One composite number: confidence-weighted value over AI cost, quarterly, baselined before the build.
6. What are these value estimates built on? Label every figure measured, estimated, or projected, with the weights disclosed. Boards forgive uncertainty; they do not forgive unlabeled uncertainty.
On execution risk
7. Who owns this? A named executive with P&L authority, and the CEO's calendar visibly attached to the first workflow. Programs owned by committees get abandoned the first busy quarter.
8. What happens to the team? Bring the role-by-role workforce plan: cost savings named honestly, redeployment planned role by role. Boards accept honest reductions. What they punish is finding out later.
9. What if it disrupts the running business? The phased sequence, people making the judgment calls, and a parallel build for the workflows that cannot afford disruption.
On the alternative
10. What happens if we wait a year? The survival math: the category is being repriced at some rate, and the compounding gap between movers and waiters is the real cost of the pause.
11. Why not just buy tools and stay flexible? That is the strategy the failed 95% ran. The flexibility is real, but the measured return from that strategy was nothing.
12. What would make us stop? State the stop rule before they ask: if the composite misses the bar two consecutive quarters, the program stops. The program that can be stopped is the program that gets funded.
The full board pack, a seven-slide structure with these twelve questions and fundable answers, is in The 2026 AI Strategy, free. Walk in with the answers written down, and the board meeting becomes what it should be: a funding decision, not an ambush.
