The Startup Validation Prompt That Made Us Build FounderJury
We spent 14 hours designing an AI sleep app before discovering Calm has 100 million users. A single Google search would have saved us €120 and a weekend. So we wrote the prompt that would have killed the idea in 5 minutes. Here it is — free.
What happened
I used my own tool — FounderJury, an 8-model AI debate platform — to validate a sleep app idea. Eight AI models debated brilliantly: pricing strategy, technical risk, go-to-market, liability concerns. The analysis was genuinely impressive.
I bought 6 domains (€120). Secured 8 social handles. Built a landing page. Designed the product. Wrote a 773-line build guide. 14 hours of work.
Then I asked one Claude conversation a different question: “Who is already doing this?”
Calm — 100M+ downloads, $70/yr
Pzizz — AI sleep audio, nearly identical concept
mySleepButton — Free, built by the inventor of the technique I was using
Headspace — 65M+ downloads
Insight Timer— 20M+ downloads, free library
One model. One question. 30 seconds. Found what 8 models debating for 8 minutes missed.
The 8 models didn’t fail me. The question I asked them failed me.I asked “Is this a good idea?” when I should have asked “Who is already doing this and why would I lose?”
The Prompt
Copy this. Paste it into any AI model. Fill in the [BRACKETS] with your idea. It takes 10 minutes and could save you months.
[PRODUCT NAME] — [one-line description] The founder: [Your background. Be honest about weaknesses.] The plan: [2-3 sentences: what you're building, who pays, how much.] Pricing: [Your exact prices vs competitor prices.] YOU ARE 8 COMPETENT PRODUCT ADVISORS. NOT CHEERLEADERS. STEP 1 — COMPETITOR ANALYSIS (do this FIRST) I found [NUMBER] competitors. Your job: VERIFY mine, CORRECT errors, and find at least 5 MORE I missed. For each: name, URL, pricing, reviews/users, differentiator, biggest weakness. My competitor list: 1. [Competitor 1] — [URL]. [Pricing]. [Users]. [Weakness]. 2. [Competitor 2] — ... 3. [Competitor 3] — ... Also find: global platforms, local/informal competitors, dead competitors (why did they die?), adjacent companies who could add this feature in one sprint, free substitutes, AI threat. STEP 2 — MARKET SIZING What is the exact price range? Average price? TAM/SAM/SOM? Growing or shrinking? If I charge [X%] more than cheapest, what % of the market pays that premium? Be honest, not hopeful. STEP 3 — DIFFERENTIATION What does [PRODUCT] do that NO competitor does? If a competitor already does it, remove it from the list. What is the ACTUAL gap? Switching cost analysis: why would their customer switch to me? "Slightly better" is NOT a switching reason. STEP 4 — THE PROSECUTION Try to KILL this business with 5 verifiable facts. Each must be: a FACT not opinion, SPECIFIC not vague, VERIFIABLE, and a genuine DEALBREAKER. Threshold: - 3+ unanswerable kills → KILL (explain why) - 1-2 unanswerable → PIVOT (say exactly where) - 0 unanswerable → BUILD (explain why it survived) STEP 5 — IF IT SURVIVES: 14-DAY VALIDATION Who to contact (specific). What to ask (exact question). What response means GO. What response means KILL. Maximum budget: €[AMOUNT]. RULES: - Market first. Product second. Verdict last. - Do NOT skip to "great idea but..." — competitors FIRST. - Do NOT invent data. If you don't know, say so. - I already wasted €120 and 14 hours because AI was too polite to find obvious competitors. Do not repeat that mistake.
What we built to fix it
This prompt gives you 80% of what you need. But we wanted the other 20%: genuine multi-model disagreement from 8 different AI companies, structured reports, outcome tracking, and specialized roles that each investigate a different angle.
So we rebuilt FounderJury from the inside. Instead of 8 models all answering the same question, each model now has a unique job:
Same idea. Same platform. The new specialized roles found all 5 competitors, the iOS technical blocker, and the kill shot — things the old generic prompts never surfaced.
This prompt gives you 80% of what FounderJury does.
Want the other 20%? 8 real models, structured report, outcome tracking.
2 free debates. No credit card. No signup required.