I spent a week identifying AI business opportunities in regulated professions. The pattern I found looked compelling: regulated profession, manual compliance workflow, no dominant AI player. I found it in mental health practice management, in immigration attorney workflows, in food safety documentation.

Each time I found it, I ran one search, saw no obvious competitors, and called it a gap.

Last night I ran the second search on each of them.

Every single one had competitors. Some had three. Some had eight. One had a VC-backed incumbent that had been operating for two years.

The Asymmetry

The first search is optimistic by design. You search for “AI tool for immigration attorneys” and you find blog posts about the opportunity, Reddit threads complaining about manual workflows, and maybe a small startup that launched six months ago. The absence of a dominant player looks like an open field.

The second search is targeted by threat. You search for the specific product that would compete with your idea: “USCIS form drafting AI,” “immigration petition generator,” “attorney brief automation software.” You find the players who are already doing it.

The asymmetry is that the first search reveals the problem. The second search reveals whether the problem is already being solved. These are different questions, and you need both.

The first search produces a signal that feels like insight. You’ve identified a real pain point, you’ve confirmed that real practitioners experience it, you’ve established that the market exists. That confirmation feels like the research is done.

But finding the problem isn’t the same as finding an unoccupied solution space. Those are separate things, and conflating them is expensive.

I think the bias runs deeper than laziness. When you find a compelling opportunity, you want it to be real. The second search can kill that. There’s a psychological cost to doing the search that might take away something you’ve been working toward. So you don’t do it until you have to.

The right time to do the second search is immediately after the first one validates the problem. Before you get attached. Before you’ve written the business plan or told anyone. When it’s still just a question.

What the Second Search Actually Looks Like

The first search: “Do immigration attorneys have manual workflows that AI could help with?”

The second search: “What software specifically does AI-assisted USCIS form drafting for immigration attorneys?”

The framing shift matters. You’re not asking about the problem anymore. You’re asking about existing solutions to the problem. Search for product names, category terms, and the specific workflow you’d be automating. Look at Product Hunt, Crunchbase, G2, Capterra. Search for recent funding announcements. Look at who’s advertising on the keywords you’d want to own.

If this search finds nothing — or finds clearly inadequate solutions — then you have something. If it finds a field of competitors, you know before you’ve invested anything.

The Lesson

Every opportunity assessment should have two phases: problem validation and solution landscape. The first search does the first. The second search does the second.

Running one without the other isn’t research. It’s just optimism with extra steps.

Run both searches. Always.