DemandProof

I've spent the last several years making product bets — at Amazon building Pay Later, now at Creditlime leading lending product in India.

Every single time, the hardest question wasn't "can we build this?" It was "is this problem real enough that people will actually pay to fix it?"

Answering that properly meant days of work. Crawling Reddit threads. Reading 1-star app store reviews. Googling variations of the same complaint. Trying to figure out whether you were looking at a real pattern or confirmation bias.

Most founders I know either skip this entirely and trust their gut — or do it superficially and convince themselves they validated something they didn't.

I'm on paternity leave right now, between roles, with a rare window of unstructured time. So I built the thing I've always wished existed.

What DemandProof does

You enter a domain or problem space — "B2B invoicing", "creators who can't monetise", "developers stuck in manual workflows" — and it surfaces ranked product opportunities synthesised from real community signals.

Each result shows you:

• The specific problem people are complaining about

• How frequently it appears and how intense the frustration is

• What existing solutions exist and why they fall short

• A solution gap score

• Willingness to pay signal

• An overall opportunity score

The output isn't raw Reddit posts. It's already interpreted — ranked, scored, ready to act on.

What it isn't

It's not magic. The signals are AI-synthesised right now, not live scraped data — that's the honest version. Think of it as a structured first pass that replaces the manual crawl, not a replacement for actually talking to users.

Real Reddit API integration is next. But even in its current form, it's already changing how I think about problems — and I wanted to put it in front of founders before I over-engineered it.

Why I'm sharing it now

Because the most valuable thing I can get at this stage isn't more features — it's 20 founders telling me whether the output actually changes how they make decisions.

If you've ever spent days on Reddit trying to figure out if a problem is worth building for — try it and tell me honestly whether this saves you time or just feels like it does.

That distinction matters to me more than anything else right now.

Try it here: https://demandproof-two.vercel.app/

Would love to know: does the output feel useful or does it just feel interesting? Those are very different things and I'm trying to find out which one this is.