How do I keep brainstorming from becoming an endless supply of unfinished ideas?
AI makes it easy to keep a brainstorm going long after the team has enough options. Decide ahead of time what the session should help you choose, and who does the choosing. An exploratory meeting can end with a question. A selection meeting needs a selection.
Tyra Banks distinguished having ideas from executing them. A session can run on with nobody deciding anything, and AI makes that easier.
Before you start, define the constraints that matter and name the person responsible for choosing. Afterward, compare the strongest options against the same criteria: what each would require, which assumptions are still open, and what the team would have to stop doing to pursue it.
Park the rest on a deferred list with no implied commitment. Resist describing every option as exciting work the team will somehow find time to deliver.
End with the selected direction, the first deliverable and the next review. Write down why the alternatives lost, so the next meeting doesn't reopen the same argument without new evidence. I'd judge a brainstorm by whether people leave with fewer unresolved choices.