AI Belongs on the Board Agenda
Lately, one of the most frequent questions I get from board members and leaders is: How should we be thinking about AI?
AI is here and much discussion swirls around the risks. Boards are moving quickly to write policy. But is policy what we need? Certainly, if there are clear risks attached to AI in relation to your mission or your operations, some philosophy or guidelines matter. AI is changing so quickly that any policy is likely to need frequent review. But what's more important is that the obvious risks you want to mitigate with policy are likely not representative of the more important dimensions of AI and its implications — both as a powerful engine and as a potential threat. AI is a much bigger conversation that will not hold still.
I was interested to read Protiviti's 2026 Global Board Governance Survey. The research indicates only about a quarter of directors reported that their boards discuss AI at every meeting. More telling is what separated the rest. The boards that treated AI as a standing matter — a recurring question rather than a settled one — were far more likely to report that they are extracting genuine value from it, while those that addressed it once and looked away were not. So boards must approach AI not as a risk and policy matter and instead as a strategy question, and strategy questions are never entirely closed.
I would offer this distinction to any board chair right now: A policy is an artifact; governance is a practice. The most important thing a board can do about AI is not to permit it or ban it in a single sitting, but to keep returning to the right questions as the tools, the risks, and the organization's own use of AI continues to change. Rather than assembling a team to write a single policy, what if instead you assembled a team to explore use cases, and develop a philosophy and guidelines?
It’s important to note the surveys I am drawing on come from the corporate world, where directors answer to shareholders and regulators. Nonprofit and independent school boards govern differently — and carry different risks. But the underlying habit — treating a fast-moving capability as ongoing work rather than a one-time rule — translates directly, and schools and not-for-profits can ill afford to learn it late.
So what are those questions? Here are a few to put on the agenda:
How can AI be harnessed for value creation, to advance our mission? Boards instinctively ask whether AI is safe and compliant, and those questions matter. But the bigger questions circle around how the organization is beginning to use these tools to improve mission delivery, and are those ways consistent with what it exists to do. A development office that lets a model decide who to cultivate, or a school that quietly outsources judgment for assessing student work, may be foregoing what humans do best and drifting from its mission. On the other hand, a development office that uses AI to make its technical work more productive and efficient, to sort and analyze data with ease, and to partner in building personalized strategies, or a teacher using AI to develop projects and learning experiences where students must also learn to use AI well — that’s a different story. A board must develop a point of view on how AI is to be utilized, and how it is not.
How can we use AI to make us more productive without outsourcing our brains? The second is about judgment. The work that AI handles well — drafting, summarizing, sorting — is real, and it frees people for more valuable work. But the work that defines any mission-driven organization is judgment: deciding between competing goods, weighing the needs of the people it serves, choosing what it will not do. Boards need to be clear about which capabilities can be handed off and which it intends to protect, and to say so plainly.
What do we know, what do we not know, and what remains unknowable? This is all about the board's own readiness. A board cannot govern what it does not understand, and AI has widened the gap between the trustees who follow this closely and those who would rather not. That gap is itself a governance risk. It does not require every trustee to become an expert. It does require the board to build enough shared literacy to ask informed questions and recognize a weak answer — rather than defer, gratefully, to whoever in the room seems most fluent.
Let me be clear: None of this calls for alarm, and none of it calls for another document. But going forward, boards need to treat AI the way they treat finance or strategy: as a living, dynamic responsibility that is always on the radar screen. Task forces and a two-page policy can be helpful, but they are beginnings rather than a conclusion.
So as boards head into summer planning, I encourage you to ask this question: When does the board next intend to talk about AI — and what will you want to know when you do?
Greenwich Leadership Partners works with nonprofit and independent school boards and leaders on strategy, governance, and leadership.