Use AI for Brainstorming, Then Narrow With Real Criteria

AI is good at giving me options quickly. It is less useful if I let it decide which option is best without knowing my constraints. I split brainstorming into two stages. If your shortlist is still too broad, Ask for Alternatives, Then Choose With a Rubric gives you a rubric for choosing between alternatives.

For a game challenge, I might ask: Give eight challenge formats that let players show off a route choice. Vary what entrants submit and how viewers judge. Do not rank the ideas yet. Keep each to two sentences. I inspect whether the formats are actually different. If six are “fastest route” with new names, I ask for alternatives built around creativity, consistency, risk, or explaining the decision.

I write criteria before ranking: can entrants record it easily, can voters understand it, is cheating obvious, does it require the same game/map, and is it fun to watch? Then I ask AI to compare the ideas against those criteria and flag where evidence is missing. This may show that “best route” and “fastest time” need different judging rules. One is an audience choice; the other needs a fixed start/end and verified timing. The distinction comes from the criteria, not from a catchy title.

Before publishing, show the rules to a few people. Ask them what they think they would submit and how they would vote. If they misunderstand, revise the rules. AI can draft options; people reveal confusion. Try it: Brainstorm ten ideas, choose three criteria, and narrow to two. Test those two on a real person or a small prototype before committing to one. Keep learning: Ask for Alternatives, Then Choose With a Rubric, Prompt AI for a Game Idea Without Getting Generic Tropes, and How to Tell Whether an AI Answer Is Any Good.