12 Questions Your Board Should Ask About AI (And How to Answer Them)
Created on 2026-02-06 09:38
Published on 2026-03-09 10:00
A preparation guide for the conversations that determine whether AI initiatives survive or die
The board meeting is scheduled. AI is on the agenda.
You know the questions are coming. The board has been reading about AI in every business publication. They have seen competitors announce initiatives. They have heard from consultants, vendors, and fellow directors about AI transformation.
They are going to ask about your AI strategy.
The quality of your answers will determine more than the outcome of this meeting. It will determine whether you get resources or resistance. Whether initiatives are protected through the valley or killed at the first sign of difficulty. Whether the board becomes an asset or an obstacle.
Most executives prepare for the wrong questions. They prepare to explain technology, to present timelines, to justify budgets.
Boards do not actually care about these things. Not primarily.
Boards care about strategic positioning, risk management, and organizational capability. They ask about AI because they want to understand these deeper concerns.
This article provides the twelve questions sophisticated boards ask about AI, and how to answer them in ways that build confidence rather than create anxiety.
The Questions Boards Should Ask
Some boards ask surface questions. License costs. Implementation timelines. Vendor selections.
Sophisticated boards ask deeper questions. These are the questions that reveal whether AI initiatives are truly ready or merely approved.
If your board is not asking these questions, consider provoking them yourself. It is better to surface hard questions before investment than to face them after failure.
Question 1: What is our AI thesis?
What the board is really asking:
Do we have a coherent view of what AI means for this business? Is there a strategic logic, or are we just doing AI because everyone else is?
What a weak answer sounds like:
“AI will help us be more efficient and innovative across the organization. We are exploring multiple use cases to capture value.”
This answer says nothing. It could apply to any technology. It reveals no strategic thinking.
What a strong answer sounds like:
“Our AI thesis is specific to our competitive position. We believe AI will fundamentally change how customers evaluate providers in our industry. The winners will be organizations that can personalize service at scale while reducing cost to serve. We are building AI capability to achieve personalization that our competitors cannot match because they lack our customer relationship data.”
This answer is specific. It connects AI to competitive dynamics. It articulates why you can win.
How to prepare:
Develop a two-sentence AI thesis before the meeting. It should answer:
What competitive dynamic is AI changing in our industry?
Why are we positioned to win in that changed environment?
If you cannot articulate this, you are not ready for the board conversation.
Question 2: What happens if we do nothing?
What the board is really asking:
Is AI optional or existential? Can we wait and see, or is delay itself a decision with consequences?
What a weak answer sounds like:
“We would fall behind competitors. We need to stay current with technology trends.”
This is generic fear without specificity. It does not answer whether “falling behind” matters materially.
What a strong answer sounds like:
“If we do nothing for 18 months, three specific things happen. First, our largest competitor, which is already deploying customer service AI, will establish service cost advantages of approximately 15-20% that we will struggle to match. Second, customers will develop expectations based on AI-enabled competitors that our human-only service cannot meet. Third, AI talent that we could recruit today will have committed to competitors. The cost of inaction is not stasis. It is competitive deterioration.”
This answer quantifies consequences. It names specific competitors. It establishes a timeline.
How to prepare:
Analyze what happens if you wait 12 months, 18 months, 24 months. Be specific about competitive consequences. Name competitors. Estimate impacts.
Question 3: Why will our AI initiatives succeed when most fail?
What the board is really asking:
Do you understand why AI fails? Have you addressed those factors? Or will you repeat the industry-wide pattern of failure?
What a weak answer sounds like:
“We have an experienced team and good technology partners. We are following best practices.”
This answer demonstrates no understanding of why 95% of AI initiatives fail.
What a strong answer sounds like:
“MIT research shows that 95% of organizations get zero return from AI investments, and the determining factor is not technology quality but organizational approach. Most organizations fail because they deploy AI into misaligned leadership, inaccessible data, undeveloped capability, broken processes, and unclear governance. We have assessed ourselves across all six dimensions of AI readiness. Our gaps are in data accessibility and process documentation. We are addressing these before deployment, not during deployment. This is why we expect to be in the 5%, not the 95%.”
This answer shows you understand why failure happens. It shows you have assessed your situation. It shows you are addressing gaps deliberately.
How to prepare:
Know the MIT research. Conduct honest readiness assessment. Identify your specific gaps. Articulate how you are addressing them.
Question 4: What is the leadership team’s level of AI understanding?
What the board is really asking:
Is the CEO going to lead this transformation, or delegate it to IT? Do executives understand AI enough to make good decisions?
What a weak answer sounds like:
“Our CIO is leading AI strategy. We have also hired a Head of AI with strong technical credentials.”
This answer confirms delegation, which is a warning sign.
What a strong answer sounds like:
“Every member of the executive team can articulate how AI will change their function. The CFO understands how AI affects financial modeling and the J-Curve of AI investment. The CHRO understands the capability development required. The COO understands the process redesign needed. I personally use AI tools daily for my own work. We are not delegating AI to IT. We are leading it from the executive team.”
This answer demonstrates that AI is understood across leadership. It shows personal adoption. It shows distributed ownership.
How to prepare:
Ensure executive team members can discuss AI substantively for their functions. Demonstrate your own personal adoption. Show that AI understanding exists across leadership, not just in technical roles.
Question 5: What does success look like at 90 days, 180 days, and one year?
What the board is really asking:
Do you have realistic expectations? Can you measure progress? Or will this be an endless initiative with no accountability?
What a weak answer sounds like:
“We expect to see significant value from AI within 12-18 months as adoption spreads across the organization.”
This answer has no specifics. It cannot be measured. It avoids accountability.
What a strong answer sounds like:
“At 90 days, we will have deployed our first AI application to the customer service team. Success means 60% of eligible interactions assisted by AI, with customer satisfaction maintained or improved. At 180 days, we will have expanded to two additional use cases and developed the Auditor Mindset in 40% of users. At one year, we will have AI embedded in five core processes, measurable productivity improvement of 15%, and institutional capability to deploy AI to new use cases without external support.”
This answer provides specific milestones. It can be measured. It creates accountability.
How to prepare:
Define specific success metrics for each time horizon. Make them measurable. Be realistic, accounting for the J-Curve.
Question 6: What will go wrong?
What the board is really asking:
Do you understand the risks? Are you planning for problems, or assuming everything will work?
What a weak answer sounds like:
“We have engaged experienced partners and are following proven methodologies to minimize risk.”
This answer pretends risk can be eliminated. Boards know better.
What a strong answer sounds like:
“Several things will likely go wrong. First, productivity will decline before it improves. Research shows this J-Curve is normal, with declines of 20-40% before recovery. We have set expectations accordingly and built a six-month runway before we expect positive metrics. Second, data quality issues will emerge that we have not anticipated. We are budgeting for remediation. Third, some users will resist adoption. We are planning change management but expect pockets of resistance. Fourth, our first deployment will have errors that we will learn from. We are starting with lower-stakes applications where errors are survivable.”
This answer demonstrates understanding of what goes wrong. It shows planning for problems. It builds credibility through honesty.
How to prepare:
Identify the most likely problems. Plan for them. Present them honestly. Pretending risks do not exist destroys credibility.
Question 7: How will we protect this initiative when results look bad?
What the board is really asking:
Do you understand that premature evaluation kills initiatives? Do you have a plan for the valley?
What a weak answer sounds like:
“We are confident in our approach and expect positive results.”
This answer ignores the reality of the J-Curve. It sets up failure.
What a strong answer sounds like:
“The J-Curve means results will look bad before they look good. We are protecting the initiative in three ways. First, we are setting expectations now, before investment, that productivity improvement will not appear until approximately month nine. Second, we will measure and report leading indicators, adoption rates, capability development, user engagement, that predict success while lagging indicators are still negative. Third, we are asking for explicit commitment from this board to evaluate the initiative at month twelve, not month six, unless leading indicators turn negative.”
This answer shows you understand the J-Curve. It shows you are managing expectations. It requests specific board commitment.
How to prepare:
Understand the J-Curve. Plan leading indicators. Request explicit commitment to appropriate evaluation timing.
Question 8: What is the total investment required, including what is not in this budget?
What the board is really asking:
Are you showing us the real cost, or just the visible costs? Will we face budget overruns?
What a weak answer sounds like:
“The initiative requires $2 million for platform licensing and implementation.”
This answer shows only visible costs. The board knows there is more.
What a strong answer sounds like:
“The total investment has four components. First, direct costs of $2 million for licensing and implementation. Second, data readiness investment of approximately $400,000 to address accessibility and quality gaps. Third, capability development costs of $300,000 for training that goes beyond tool usage to judgment development. Fourth, productivity dip during the J-Curve, which we estimate at $500,000 in reduced output during the learning period. Total investment is approximately $3.2 million, not $2 million. We have built 25% contingency on top of this.”
This answer shows complete cost understanding. It separates visible and hidden costs. It builds credibility through honesty.
How to prepare:
Calculate total cost of ownership, including hidden costs. Separate direct costs, data costs, capability costs, and productivity costs. Add contingency.
Question 9: How does this change our risk profile?
What the board is really asking:
What new risks are we taking on? How do we manage them? Are we creating liabilities?
What a weak answer sounds like:
“We are working with our legal and compliance teams to ensure appropriate risk management.”
This answer defers rather than addresses. Boards want substance, not process.
What a strong answer sounds like:
“AI changes our risk profile in four ways. First, we create data security exposure. AI systems access more data across more systems. We are implementing data classification and access controls appropriate to AI usage. Second, we create accuracy risk. AI will make errors. We are implementing human review for high-stakes outputs and building the Auditor Mindset so users catch errors. Third, we create bias risk. AI can produce discriminatory outcomes. We are implementing testing for bias in customer-facing applications. Fourth, we create dependency risk. If AI systems fail, can we operate? We are maintaining human capability for critical processes.”
This answer identifies specific risks. It describes specific mitigations. It shows governance thinking.
How to prepare:
Map the specific risks AI creates. Develop specific mitigations. Be prepared to discuss each category.
Question 10: What is our governance framework?
What the board is really asking:
Do we have appropriate oversight? Is accountability clear? Can we defend our decisions if something goes wrong?
What a weak answer sounds like:
“We are developing policies and will have a governance committee in place.”
This answer is aspirational. Boards want to know what governance exists now.
What a strong answer sounds like:
“Our governance framework has four components. First, a clear decision rights matrix: who can approve what AI applications for what uses. The board approves applications above a risk threshold we have defined. The executive team approves moderate-risk applications. Business unit leaders approve low-risk applications. Second, accountability is specific: each AI application has a named accountable executive, not a committee. Third, we have policies for data usage, human review, and error handling that are documented and trained. Fourth, we have an audit mechanism: quarterly review of AI performance and issues.”
This answer provides specific governance structure. It shows clear accountability. It demonstrates thinking has been done.
How to prepare:
Develop governance framework before board presentation. Define decision rights. Assign accountability. Document policies.
Question 11: What is our competitive position relative to peers?
What the board is really asking:
Are we leading, following, or falling behind? How does our AI maturity compare?
What a weak answer sounds like:
“We are monitoring competitor announcements and are on track to remain competitive.”
This answer provides no information. Boards want specifics.
What a strong answer sounds like:
“Based on our analysis, we are in the middle of our peer group. Competitor A has deployed customer service AI and claims 20% cost reduction. Competitor B has announced significant AI investment but visible deployment has been limited. Competitor C appears to be in early stages similar to us. We are not leading, but we are not significantly behind. Our approach of building readiness before deployment means we may deploy later but with higher probability of success.”
This answer provides specific competitive positioning. It demonstrates market awareness. It explains your strategic rationale.
How to prepare:
Research competitor AI activities. Assess their maturity versus yours. Position your approach relative to theirs.
Question 12: How will we know when to stop?
What the board is really asking:
If this is not working, will you tell us? Do you have criteria for killing the initiative? Or will you keep spending regardless?
What a weak answer sounds like:
“We are committed to making this work and will adjust as needed.”
This answer suggests sunk cost thinking. Boards want to see objectivity.
What a strong answer sounds like:
“We have defined kill criteria. If at month six we see declining adoption, negative quality indicators, and unresolved blockers, we will recommend pausing for reassessment. If at month twelve we have not achieved the leading indicator targets we defined, we will recommend significant restructure or termination. We have committed to intellectual honesty about whether this is working. The worst outcome is continued investment in a failing initiative. We will not let that happen.”
This answer shows you can be objective. It defines specific kill criteria. It demonstrates stewardship, not advocacy.
How to prepare:
Define kill criteria in advance. What would make you recommend stopping? Be specific.
How to Handle Questions You Cannot Answer
Sometimes boards ask questions you cannot answer. This is normal. How you handle it matters.
Do not bluff.
Boards can detect bluffing. It destroys credibility. If you do not know, say so.
“I do not have that data. I will provide it after this meeting.”
“That is a good question we have not fully worked through. Let me come back with an answer.”
Redirect to what you do know.
“I cannot speak to that specific comparison, but here is what we do know about our competitive position…”
“That detailed analysis is not complete, but at a directional level, we believe…”
Commit to follow-up.
“I will provide a written response by Friday on that question.”
Then actually follow up. Reliability builds credibility over time.
Preparing for the Conversation
Beyond preparing answers to specific questions, prepare for the conversation itself.
Know your board members’ concerns.
Different board members have different priorities. The member from a technology background will ask different questions than the member from a financial background.
Anticipate what each member cares about. Prepare answers that address their specific concerns.
Prepare for follow-up questions.
Your first answer will rarely be the last word. Prepare for follow-up questions that go deeper.
“You mentioned data readiness is a gap. What specifically is the gap and how long will it take to close?”
“You said productivity will decline first. How steep a decline and for how long?”
Have supporting detail available.
Bring supporting material.
Have backup slides or data available even if you do not plan to present them. When questions go deep, you can provide supporting material.
“Here is the detailed timeline if you would like to see it.”
“This slide shows our competitive analysis in more detail.”
Practice the difficult questions.
The questions you hope they will not ask are the questions you need to practice most. Have colleagues ask you the hardest questions. Practice answering them out loud.
Boards determine whether AI initiatives get resources, protection, and patience.
The quality of your board conversation determines whether you have an asset or an obstacle. The twelve questions in this article are the questions that matter. Prepare for them.
Answer with specificity, not generality. Answer with honesty about risks, not pretense of certainty. Answer with evidence of thinking, not just claims of competence.
Boards that understand AI transformation can support it. Your job is to help them understand.
What questions has your board asked about AI? What questions should they be asking?
The AI Readiness Scorecard provides the honest assessment that prepares you for board conversations. It takes ten minutes and reveals the gaps you need to be prepared to discuss.
Comment “SCORECARD” below and I will send you access.
The board conversation is coming. The question is whether you will be ready for it.