Most companies with a stated AI strategy hold plenty of meetings about executing it, and comparatively few dedicated meetings whose sole purpose is comparing that strategy against what has actually happened since it was set, with the explicit mandate to revise the strategy if the evidence warrants it.
Why this specific meeting rarely happens
Execution-focused meetings have an obvious agenda and clear outputs; a review meeting whose purpose is potentially concluding that the current strategy is wrong is a harder meeting to justify calendar time for, especially when no specific crisis has prompted it. It's the kind of meeting that's easy to indefinitely postpone in favor of more immediately actionable work.
What a real review meeting requires
Comparing specific predictions made when the strategy was set against actual outcomes, with someone in the room specifically tasked with arguing for revision rather than defending the existing plan by default. Without an explicit devil's advocate role, review meetings tend to drift into confirming the existing strategy regardless of the evidence, simply because no one in the room is incentivized to argue otherwise.
Why this matters more for AI strategy specifically
AI capability and cost curves move fast enough that a strategy set even a year earlier can be built on assumptions that no longer hold, in either direction - a capability gap that's since closed, or a cost assumption that's since become outdated. Without a scheduled review, these shifts go unaddressed until they cause a visible problem.
A simple fix
Scheduling this review on a fixed calendar cadence, with a defined agenda and an assigned skeptic, converts a meeting that's easy to skip into one that happens by default - a small structural change that meaningfully increases the odds a company notices when its own strategy has stopped matching reality.