AI is forcing boards to play offense and defense at the same time.
On one side is opportunity: new business models, productivity, speed, customer experience, competitive advantage and entirely new sources of enterprise value.
On the other is risk: security, regulation, reputation, workforce implications, model behavior, accountability and decisions increasingly made or influenced by machines.
So the question I hear more and more is a natural one:
How do we get the right AI expertise into the boardroom?
The reflexive answer is often: Appoint an AI expert.
Sometimes (but rarely) that is exactly right.
But I think the question is more nuanced than that.
A board seat is not a short-term advisory assignment. It is a fiduciary role across the entire enterprise agenda. Directors are responsible not only for AI, but strategy, capital allocation, succession, risk, compensation, culture, financial performance and the long-term health of the company.
So before filling a scarce board seat with a deeply technical AI specialist, I would ask a different question:
What kind of human capability does this particular board need in order to govern AI well?
I see at least four models.
(And, as anyone who knows me, I am highly suspicious of one-size-fits-all.)
1. Appoint the AI expert
Let's start with the obvious one.
For some companies, a true AI expert belongs on the board.
If AI is fundamental to the business model, the product, the competitive moat or the company's future economics, deep technical expertise can be extraordinarily valuable.
Amazon's appointment of Andrew Ng is a useful example. Ng brings decades of work in machine learning, AI education, entrepreneurship and AI-company building. Amazon explicitly identifies his AI experience as informing the board's perspective on AI's opportunities, challenges and transformative business potential.
For the right company, that makes complete sense.
But I would resist turning that example into a universal prescription.
A world-class AI scientist may bring extraordinary depth on the technology and still not be the best use of a permanent fiduciary seat for every company.
And there is another practical issue: the AI leader with the greatest technical currency may not particularly want the full obligations of public-company board service. Nor should we assume that deep technical expertise automatically translates into exceptional governance judgment.
So the question is not:
Would an AI expert add value?
Almost certainly.
The better question is:
Does the value of that expertise warrant a board seat?
For some companies, absolutely.
For many others, there may be better architectures.
2. Appoint the technology-proficient value creator
This may be the more useful archetype for a much broader group of boards.
I am thinking about a leader who is technologically proficient without necessarily being an AI scientist.
They understand enough.
But more importantly, they have lived through multiple technology cycles and understand how technology moves from interesting capability to business-model consequence.
They have seen technology reshape:
- competitive advantage
- customer behavior
- operating models
- economics
- talent requirements
- organizational design
This director becomes an advocate for technology as a business lever, not technology for technology's sake.
They can ask management questions such as:
Are we investing in the right things?
Where is AI actually changing the economics of the business & industry?
What leadership changes are required to win in tomorrow’s market?
And what are our competitors learning faster than we are?
That is a different kind of AI fluency.
It is less about being able to explain precisely how the model works and more about knowing where to probe to determine whether the enterprise is actually capturing value from it.
For many boards, that may be closer to the capability they need.
3. Appoint the bold enterprise decision-maker
This one may sound counterintuitive.
What if one of the strongest additions to an AI-era board isn't particularly technical at all?
Consider a seasoned enterprise leader who has repeatedly made consequential decisions in moments when the answer was unclear.
Perhaps they have exited major markets because geopolitical conditions changed.
Maybe they redirected billions in capital from a historically successful business into an uncertain new category.
Maybe they challenged an industry's prevailing business model before the economics made the answer obvious.
Maybe they led a company through a structural disruption that forced uncomfortable choices about portfolio, workforce, operations or strategy.
What does that have to do with AI?
Potentially quite a lot.
Because governing AI will not only require understanding the technology.
It will require judgment under uncertainty.
Boards will have to decide when to accelerate and when to pause.
Where to place large bets before the evidence is conclusive.
Which legacy assumptions are becoming dangerous.
How much organizational disruption is justified by potential future value.
And when the greater risk is no longer moving too fast, but moving too slowly.
I would not discount the leader who has demonstrated intellectual rigor, courage and judgment simply because AI wasn't the context in which they developed those muscles.
In fact, that person may bring exactly the pattern recognition the board needs.
The real question becomes:
Has this leader demonstrated an ability to recognize a consequential inflection point and act before the path became obvious?
That may be one of the most important governance capabilities of the AI era.
4. Build an AI receptor into the boardroom
The fourth model is different because it separates two things we often try to solve with one person:
governance accountability and technical currency.
AI is moving exceptionally fast.
Alpha's work on continuous governance makes the challenge clear: systems, capabilities and operating realities can change materially between traditional board meetings. Static governance mechanisms are increasingly being asked to oversee technology that does not behave statically.
That makes me wonder whether every board needs not simply an AI expert, but an AI receptor.
By receptor, I mean a clearly designated director who owns the responsibility for staying close to the AI agenda.
This may be a new director or someone already sitting on the board.
The individual needs curiosity, commitment and enough fluency to engage credibly. They should care deeply about the topic and be willing to invest real time in it.
But they do not have to personally contain every form of AI expertise. Instead, I would surround that director with a small, external AI Advisory Council.
The council could include complementary perspectives such as:
- technical AI expertise,
- cybersecurity,
- regulation and policy,
- operating-model transformation,
- human capital,
- ethics and reputation,
- and sector-specific AI applications.
I would also make the council deliberately renewable, given the speed and seismic nature of AI’s evolution. Perhaps members commit for 12 to 24 months, with rotation built into the model.
Why?
Because currency matters.
The expert who was indispensable three years ago may not represent the expertise the company needs three years from now.
The board receptor provides continuity.
The advisory council provides currency.
On a regular cadence, the receptor engages this group, pressure-tests assumptions, understands emerging issues, brings company-specific questions into the discussion, and then translates the most material insights back into the boardroom.
This creates leverage without asking one board director to be all things AI.
It also avoids using a permanent board seat to solve what may partly be a rapidly changing expertise problem.
I find this model particularly interesting because the board does not surrender accountability to outside experts.
Quite the opposite.
It creates a more intentional mechanism for improving the quality of the board's judgment.
The real question is not "Do we have an AI expert?"
I think boards could easily fall into a familiar governance trap here.
We add someone with the right credential.
We create a committee.
We put AI on the agenda.
And we feel more governed.
But representation is not the same as capability.
Having an AI expert in the room does not automatically mean the board is asking better questions, receiving better information, making better decisions or governing the enterprise's AI exposure more effectively.
The talent question has to follow the business question.
How fundamental is AI to our value creation strategy?
Where is our greatest exposure?
How quickly is the environment around us changing?
What capability already exists on this board?
Where are our blind spots?
Do we need deep technical expertise, technology-enabled business judgment, proven decision-making under disruption, or some combination?
And which capabilities need to reside permanently on the board versus remain connected to the board through a more dynamic structure?
Those questions will produce different answers for different companies.
They should.
AI governance needs a talent architecture
Much of the AI governance conversation understandably focuses on frameworks, controls, disclosures, regulation, risk and accountability.
All of that matters.
But governance is ultimately exercised by people.
Someone has to interpret the signal.
Someone has to ask the uncomfortable question.
Someone has to recognize when the risk profile has changed.
Someone has to understand when management's answer is technically correct but strategically insufficient.
And someone has to know when the greatest governance failure would be preventing the company from moving boldly enough.
That is why I increasingly think AI governance is also a board talent and architecture question.
The objective should not be to collect the most impressive AI credential available.
It should be to assemble the combination of expertise, judgment, curiosity, courage and current intelligence that this particular enterprise needs.
For one board, that may mean an Andrew Ng (Founder of DeepLearning.AI).
For another, it may mean a technology-fluent operator who has navigated several waves of disruption.
For another, it may mean an extraordinary enterprise decision-maker who has proven capable of placing consequential bets under uncertainty.
And for some, the strongest answer may be a director who serves as the board's AI receptor, supported by a deliberately rotating network of specialists.
There is no trophy for having the most technical board.
The measure is whether the board can govern the risk while helping the enterprise capture the opportunity.
And that is a much more interesting question.
