If I were sitting with a board right now, I would not ask the executive team how many AI tools they bought. I would not ask for a parade of demos. I would ask whether the company has an AI operating thesis, who owns it, how risk is governed, and what proof would tell us the work is becoming more than experimentation.
Boards have a difficult job in moments like this. Move too slowly and you become a brake on the future. Cheerlead without discipline and you become a permission slip for chaos. AI deserves neither panic nor theater. It deserves better questions.
What boards should actually ask
Start with ownership: who is accountable for AI strategy at the executive level? If the answer is scattered across departments, the company is not leading AI. It is accumulating it.
Then ask about architecture. Where does knowledge live? What data is allowed into which systems? How does the company distinguish a tool from infrastructure? How will intelligence compound inside the business instead of disappearing into vendor accounts?
Then ask about human standards. What decisions require human accountability? What customer moments should not be automated? What will happen to roles as work changes? These are not soft questions. They are governance questions.
Why board questions matter now
AI adoption is already happening inside companies whether boards have approved a formal strategy or not. Employees are testing tools. Vendors are selling urgency. Competitors are learning. Customers are forming new expectations. Silence from governance does not create safety. It creates unmanaged reality.
The board's role is not to slow AI down. It is to make sure speed has a steering wheel.
That is why I believe every CEO needs an AI strategy before the board asks for one. The best boards do not replace leadership. They sharpen it.
What fails when boards ask the wrong questions
The first failure is demo governance. Everyone watches something impressive, nods, and leaves without knowing whether it can survive legal, operational, customer, data, brand, and culture realities. A demo is not a control system.
The second failure is risk-only thinking. Risk matters. But if the board only asks what could go wrong, leadership learns to hide the future inside cautious language. Boards also need to ask what it costs to wait, what competitors can now do faster, and what customer expectations will make the current model feel outdated.
What proof I would want to see
I would want to see a written AI thesis, a workflow map, a risk policy, a training plan, a decision ledger, and proof from the business. Not vanity metrics. Not "we trained the team." Proof that the company is serving faster, deciding cleaner, preserving quality, and building institutional memory.
My own proof comes from building CI Web Group into an AI infrastructure company and learning the hard way that transformation is not one decision. It is a sequence. Rebuilding from 320 to 38 taught me that the operating model matters more than the story people want to tell about headcount.
What I would put on the next agenda
Put four questions on the next board agenda. What do we believe AI changes about our business model? Where is unmanaged AI already happening? Which workflows will we redesign first? What human standards are non-negotiable?
Then have the CEO read The Modern CEO Operating System and bring back a one-page operating thesis. A board does not need to become the AI department. It does need to make sure leadership is building the future with discipline, proof, and soul.