I kept getting “great idea” from AI, so I built something that argues back

I kept running into the same problem while using AI to think through startup ideas.

I'd describe an idea.

Then I'd ask:

"Is this a good startup?"

And I'd get a very polished response explaining why it could work.

Market opportunity.

Potential customers.

Competitors.

Pricing.

Growth opportunities.

It was useful.

But it was also a little too comfortable.

I realized I wasn't really asking AI to challenge my thinking. I was asking it to help me develop the idea I already wanted to believe in.

That became the starting point for FounderMind AI.

Instead of making the first output a score or a simple yes/no verdict, I'm building the workflow around different questions:

Business — does the model make sense?

Market — what alternatives already exist?

Product — is this actually solving a problem?

Financials — do the economics make sense?

Risk — what could make the idea fail?

The part I'm most interested in is what happens after the analysis.

Not:

"8/10. Great idea."

But:

"Here are the assumptions you still haven't proven."

That's the product I'm refining at Evolvix AI.

I'm still early, and I'm deliberately sharing the process while I figure out what is actually useful rather than pretending the workflow is finished.

If you've used AI to evaluate your own startup ideas, I'd genuinely like to know:

What do you wish the AI had challenged instead of simply agreeing with you?

5 views