AI is evolving faster than the cadence of most board meetings. That’s a governance challenge—not just for management but also for board members themselves. In my conversations with CEOs and boards, one question comes up repeatedly: How can directors use AI to become more effective governors?
It’s not about trivial uses, like improving a board pack. It’s about fulfilling the classic director’s remit: an informed, independent sense-check on the company’s policies, practices and performance. Directors should stick to enterprise-approved tools and keep confidential materials out of public AI systems. And the goal isn’t to build a parallel set of numbers or take an adversarial posture toward management. The goal is walking in with enough context to engage with management’s framing rather than simply absorb it.
That plays out across three areas: governing AI spend, benchmarking strategy and tracking regulation.
1. “How much are we really spending on running AI?”
Most boards approve AI investment at the strategy level—a budget line, a quarterly CTO update. What they don’t see is the detail underneath: compute and token spend, vendor contracts, an infrastructure bill diverging from the original case. The board approved the strategy; that’s not the same as knowing it’s being executed as intended. In many companies, total AI cost is a number only management knows.
This goes wrong fast. When infrastructure is funded centrally but consumed by individual teams, you get a textbook tragedy of the commons: No team has an incentive to contain usage because the cost lands elsewhere. Token spend scales without accountability—a governance problem, not a technology one. The fix isn’t directors building their own tracking. It’s arriving with enough grounding in vendor pricing and model economics to ask sharper questions and sense-check what management is presenting.
2. “How do we inform strategy with our own view of what competitors are doing?”
There’s an unavoidable informational asymmetry between a board and its management, which naturally frames information through the strategy it’s executing. Relying on that alone leaves assumptions untested in an area where experience and instinct can’t compensate for stale knowledge. Scanning competitor earnings calls, filings and technology announcements can now happen in hours, not days. A director could summarize every AI-related disclosure from top competitors last quarter and compare it against management’s narrative before the next meeting. Coming prepared has always been the expectation; AI just makes it achievable as the picture shifts between meetings.
3. “How do we track regulation ourselves, without waiting for briefings from management?”
Enforcement actions and compliance developments at peer companies rarely reach the board until management decides they’re relevant—by which point the framing is set. The barrier has always been time: Monitoring regulators and enforcement across jurisdictions is unrealistic for a non-executive director with a full-time job elsewhere, so it gets delegated back to management.
AI changes that math. Directors can monitor the terrain continuously—tracking enforcement and flagging what’s happening at peer companies before an issue becomes a board agenda item. That’s especially valuable in AI, where regulation moves unevenly and enforcement is starting to follow investment. Where the rules are still being written, spending time waiting to be briefed becomes a risk in itself.
Same remit, different tools
None of this changes what good oversight has always been about: independence of mind, used to ask the right questions and hold management accountable. What AI removes is the constraint of research time, so directors can stay current—even as the picture shifts between meetings.
