Request a diagnostic
The quality of board oversight depends on the quality of board questions. AI is no exception. According to the NACD 2025 Board Practices and Oversight Survey, more than 62% of directors now set aside agenda time for full-board AI discussions. The next step is making that time count.
These 25 questions are grouped into 5 areas. Use them to prepare for a management AI update, a strategy offsite or a committee review.
1. Where does AI change the economics of our business over the next 3 to 5 years: cost, revenue, pricing or competitive position?
2. Which competitors or new entrants use AI to attack our core business, and how fast?
3. What are our top 3 AI investments, and how do we measure their return?
4. Which AI pilots have we stopped, and what did we learn?
5. Are we building, buying or partnering for our most important AI capabilities, and why?
6. Do we have a complete inventory of material AI systems, with owners and risk ratings?
7. Which AI systems would we classify as mission-critical, and how are they monitored?
8. What is our AI risk appetite: where is AI encouraged, restricted or prohibited?
9. How do we detect hallucination, bias and model drift in systems already in production?
10. What is our exposure to third-party model providers, and what happens if one fails, changes terms or raises prices?
11. Which AI framework do our controls follow, such as the NIST AI RMF or ISO/IEC 42001, and has internal audit tested them?
12. Which AI laws apply to us, including the EU AI Act and state-level rules, and what is our compliance roadmap?
13. Do our public statements about AI match what our systems do? Overstating AI capability creates securities risk.
14. How do our AI risk disclosures compare with peers? The Conference Board reports that 83% of S&P 500 companies now disclose AI risk.
15. How do we describe board oversight of AI in our proxy statement, and is the description accurate?
16. Which AI incidents would we disclose, and who decides?
17. Who in management owns AI strategy and AI risk, and are those the same person?
18. How will AI change our workforce: roles, skills, headcount and hiring?
19. What AI training do employees receive, and how is it documented?
20. Do we have the AI talent we need, and can we retain it?
21. How does AI factor into executive goals and compensation?
22. Which committee owns which AI risks, and do our charters say so?
23. What AI incidents must reach the board, and how fast?
24. Does our board have enough AI fluency to challenge management? Research from MIT CISR links boards with 3 or more digitally and AI-savvy directors to stronger financial performance.
25. When did the board last assess its own AI competence, and what did we do about the gaps?
Strong answers are specific and measured. They name owners, cite metrics, describe failures as well as successes and connect AI to strategy. Vague answers ("we are exploring AI across the business") signal that management's AI program is less mature than it appears, or that the board has not asked clearly enough.
Questions only help when directors can evaluate the responses. If your board finds it hard to judge whether management's answers are strong, that is a fluency gap, not a management gap. Start with AI fluency for board directors, take the board AI self-check or request a board AI diagnostic to benchmark the full board.
At least once a year in depth at full-board level, with committee-level reporting every quarter for material AI risks.
The executive accountable for AI, often supported by the CEO, CFO, chief risk officer, general counsel and CIO or CTO, depending on the topic.
