AI Business Idea Review
Describe the idea. Four rival AIs make the case against it as hard as the case for it — and hand you a go / refine / drop verdict with the evidence.
A go / refine / drop verdict with the riskiest assumptions named.
Signed off by an independent Final QA audit — with a “verify before you rely on this” list, never a black box.
Why this team
Everyone a founder normally asks — friends, family, other founders — is structurally incentivised to encourage. This council is not. The Sceptical Investor has seen a hundred versions of the idea and asks what kills this one; the Pre-Mortem Analyst starts from “it failed — why?”; the First-Principles Thinker strips the pitch to what must be true; and the End-User Advocate asks whether the supposed customer feels the pain weekly or just nods politely. Sun Tzu weighs the terrain: who already owns the ground you want.
Each seat runs on a different frontier model — GPT, Claude, Gemini, Grok — so it’s a genuine cross-model review, not one AI in costumes. How a run works →
What the debate sounds like
An illustrative excerpt — A first-time founder wants to spend a year building a marketplace that matches home cooks with nearby customers for weeknight meal pickup.
Both sides of this marketplace have a cheaper alternative already in their pocket: the customer has supermarket ready-meals and delivery apps; the cook has a WhatsApp group and a Saturday market stall. The question is not whether the idea is nice — it is which of those two walks past their existing habit, weekly, for this.
The cold-start problem is squared here because supply is hyper-local: twenty cooks in the wrong suburb is zero supply where your first customers live. And the regulatory surface — home-kitchen food safety — is the kind of thing that does not matter until the week it shuts you down. I want the wedge narrowed to one postcode and one cuisine before a line of code.
Working backwards from failure, month eighteen: it did not die from competition, it died from repeat rate. Occasion-based food purchases spike at launch from novelty, then decay. The kill-metric to define now: if fewer than a meaningful fraction of first-time buyers reorder within a month, the habit does not exist and no amount of supply fixes it.
Questions
Will this just talk me out of my idea?
No — the format forces a verdict of GO, REFINE or DROP, and a GO from a council designed to kill weak ideas means something. What you actually get is the ranked list of assumptions that must be true, and the cheapest two-week experiment to test the fragile one — which is worth more than encouragement either way.
My idea is early and vague — is it too soon to stress-test it?
Earlier is better: the debate costs an evening, a year of building costs a year. Vague inputs produce one predictable finding — the council will tell you which part is too vague to evaluate, and that is itself the next piece of work. Founders typically re-run the brief after each refinement.
I’m worried about sharing the idea — is it confidential, and could the AI steal it?
Your brief is not used to train models and is processed only to run your council — see the security page for how sessions are handled. On the broader fear: ideas are rarely the scarce asset, execution is, and the fastest way to lose a year is refusing scrutiny of a fragile assumption.
Your material is used only to run your review — never to train public models. Encrypted in transit and at rest. Security & privacy →
Want full control — pick your own minds, set the depth? Open the full council →