The Board Doesn’t Know What to Approve
Most boards weighing AI adoption are stuck between two bad options: do nothing and risk falling behind, or approve of whatever gets pitched to them and hope it’s fine. Neither is really a decision. It’s what happens when a board has no framework for evaluating what’s in front of them — no clear read on the actual legal exposure, no sense of what a reasonable investment looks like relative to the return, and no one they trust to tell them plainly where the real risk sits versus where it doesn’t. Most boards have simply never had to evaluate a technology moving this fast with this little precedent to lean on, and it shows how these conversations go: long on concern, short on an actual answer.
That uncertainty is usually where the honest ask surfaces in conversation. Not “explain AI to us” — boards have heard plenty of that already. It’s closer to: “we don’t need another explainer, we need someone to tell us what to actually do.” That’s not a failure of diligence on the board’s part. It’s a reasonable response to being asked to govern something nobody handed them a rubric for.
Left unresolved, this uncertainty doesn’t stay neutral — it defaults to one of two outcomes, and neither is a choice the board actually made on purpose. Paralysis looks safe in the moment, right up until staff adopt AI tools on their own anyway, just without any board visibility into it. Ad hoc approval looks like progress, right up until the association has approved three different AI initiatives with no shared standard for evaluating any of them. Both outcomes trace back to the same root cause: the board was never handed something concrete enough to actually decide on.
What actually resolves this isn’t more information. Boards don’t need another AI explainer or a longer slide deck walking through the general landscape — they’ve usually already sat through one of those. What resolves it is a document the board can read in a single sitting and vote on, the same way they’d vote on any other policy that touches risk and spending:
- A plain-language risk summary, not a legal memo — what the association is actually exposed to, and what it isn’t.
- A defined approval process, so the next AI tool request doesn’t require reinventing the evaluation from scratch.
- Clear ownership, so “who’s responsible for this once it’s approved” has an answer before the vote, not after.
- A realistic cost-and-benefit framing, grounded in what the association actually does, not a generic AI ROI argument copied from somewhere else.
Consider what “clear ownership” actually looks like in practice. It’s not enough for a policy to say AI tools require approval — someone specific needs to be accountable for monitoring how those tools are actually used once the vote is over. That might mean a staff lead responsible for flagging new use cases as they come up, or a standing board committee that reviews AI-related requests on a set cadence instead of case by case. Without that person or process named in writing, “approved” quietly turns into “unsupervised” within a few months, and the board finds out only when something goes wrong. This is the difference between a policy that sits in a drawer and an AI governance policy for boards that actually gets used.
Boards don’t need to become AI experts to govern this well. They need something concrete enough to say yes or no to — the same way they’d evaluate any other policy that touches risk and spending. A board-approvable AI Governance Policy is built to be exactly that: not a slide deck about the future of AI, a document built to survive a single meeting and an actual vote.
See how it works: https://amsnow.io/services/#ai-policy
