val8

AI panel of 45 personas to validate startup ideas

September 5, 2026 Show IH: I built an AI panel of 45 personas that brutally roast startup ideas before you code

Hey Indie Hackers! 👋

Every founder has experienced this: you come up with a startup idea, pitch it to your friends or family, and they nod along saying: "That's awesome, you should build that!"

Six months and hundreds of hours later, you launch... and get total radio silence. Friends are polite. Strangers with credit cards are not.

To solve my own itch, I built val8.app an AI platform that simulates a room full of 45 realistic, critical stakeholders to test your business idea before you write a single line of production code.

How it works

Instead of one generic, polite ChatGPT prompt, val8 sets up a structured multi-agent debate across 3 distinct clusters:

1. 💼 15 Investors: Scrutinize TAM, unit economics, defensibility/moats, scalability, and exit potential.

2. 🔬 15 Subject Matter Experts (SMEs): Challenge technical feasibility, regulatory/legal hurdles, and operational bottlenecks.

3. 👥 15 Target Users: Dissect willingness to pay, switching friction, habits, and actual daily pain points.

I have specifically built skeptic preservation into the multi turn agent deliberation. There are dedicated "roasters" and "ragebait" personas who refuse to be polite and simulate your real world critics. Top finalists advance to a cross category debate and produce a final consensus report with scores (1–100) and actionable next steps.

I would love your feedback on:

Does the landing page communicate the value clearly?
If you test a concept, does the AI panel uncover blind spots you hadn't considered?
Drop your honest thoughts below!

Yash

2 Likes

3 Comments

Say something nice…

Post Comment

2

The interesting signal seems less about whether the panel finds blind spots and more about what founders do afterward. Have you seen anyone actually change a build, pivot, or pass decision because of the panel, rather than just finding the roast useful or entertaining?

Aryan Sinh

·
6 hours ago
·
Reply
1. 1
  
  This hits the exact nail on the head, and it's the #1 thing I was paranoid about when building this: avoiding becoming just an AI insult comic. Since I have just opened publicly today, my sample size is still early (about ~25 private tests will fellow founders), but I've already observed 2 distinct behavior shifts: 1. The "Smart Kill" (Passing early): One founder was planning a B2C social/discovery tool. The "User" personas consistently revolted on switching friction pointing out that while the idea sounded fun, 0 users would bother migrating off their existing Whatsapp group habits. Seeing 15 simulated users independently call out the same adoption barrier made him shelf the idea before on freelance dev work. 1. The "ICP Pivot": Another founder brought a SaaS idea. The SME cluster flagged regulatory and compliance bottlenecks they completely overlooked, while the Investor cluster tore apart their broad TAM assumptions. Instead of killing the project, they narrowed the target ICP from "all freelancers" to a very specific regulated vertical that actually has budget for it. That's why I intentionally engineered the final screen to not just give a vanity score or punchline, but: - The top objections one must solve - Structural Risks (tech, regulatory, market) - Next real world validation steps (what to ask in human interviews, what landing page test to run) Curious: from your perspective, what post validation artifact or workflow would be the most valuable bridge between the panel's critique and a founder's next action?
  
  Yash
  
  ·
  6 hours ago
  ·
  Reply
  1. 1
    
    Those are much stronger signals than people simply enjoying the critique — especially the founder who actually shelved the idea. I’d be interested in digging into that bridge from simulated objections to a real decision. Happy to continue privately — what’s the best email to reach you on?
    
    Aryan Sinh
    
    ·
    4 hours ago
    ·
    Reply

val8

About

People