AIToolRoulette

AIToolRoulette targets the genuine but moderate pain of multi-model AI routing confusion — "which AI should I use for this task?" — in a market where no single frontier model dominates every use case. The tab-switching friction is confirmed by multiple funded products (MultipleChat, AiZolo, Poe) all pitching convenience consolidation. However, the critical a16z data point tempers enthusiasm: fewer than 10% of ChatGPT users visit another model provider weekly, meaning the acute, daily pain is real only for the 10–20% power-user segment. The competitive gap is real — no consumer product auto-scores outputs and saves a personal routing card — but the addressable base of users who'd pay for *routing intelligence* (vs. just access) is narrower than the headline multi-model market implies. API cost structure at consumer price points is a real margin risk. Score improves modestly from prior 16 to 17 on stronger evidence of the routing-intelligence gap and confirmed spend patterns, but structural headwinds (91% don't multi-sub, crowded access layer, scoring credibility problem) keep this in retest-not-ship territory.