What should an AI data assistant do when two teams mean different things by the same number?

A faster dashboard will not settle an old argument about the metric. If sales calls an account ‘active’ after a meeting and finance only after payment, the assistant needs to show both definitions before it recommends a hiring plan or renewal push. For a small team, I would pilot this on one weekly question and ask the people who answer it now: did the result name the source, date range, exceptions, and definition without sending you back through three tabs? If not, the new tool has made the disagreement look more official. What would you need beside an AI-generated number before you would use it in a real team decision?