How two outsiders found a $10M+ opportunity in insurance with AI

Executive overview

Solving your own problems as a founder leads to oversaturated markets serving tech workers. The real opportunity is going where others won't look — unfamiliar industries with untouched workflows.

Jordan Katz and Michael Mattheakis entered insurance with no background, discovered massive manual accounting workflows, and built Comulate into an eight-figure ARR business.

Outsider status is an advantage: deep curiosity about unfamiliar workflows beats domain familiarity.

Finding the idea in an unfamiliar industry

• "Solve your own problems" advice concentrates founders on tech-industry problems — not where the biggest gaps are.
• Insurance revealed paper-heavy, spreadsheet-driven workflows copying data between analog and legacy digital systems.
• The specific hook: insurance has a unique accounting layer (reconciling commissions across intermediaries) with no good automation solutions.
• Initial skepticism was intense — first prospect said "we don't need your services, I'm taking this call as a favour."
• That same company is now one of their largest customers.

Validating through deep customer discovery

• Early calls focused on understanding workflows in depth — edge cases, the "why" behind each step — not selling.
• Coming in as product thinkers, not salespeople, built trust faster.
• A retiring customer noted: "50 years here — you're the first product person who ever felt like they understood us."
• You only need a few early believers; luck plays a role in finding them, but persistence determines whether you meet them.

Rethinking workflows rather than automating them

• Most companies use AI to automate existing workflows — a limited framing.
• The bigger opportunity mirrors the web-to-mobile shift: the winners reimagined what was possible, not just what existed.
• Instagram didn't replicate a website on mobile — it exploited new capabilities.
• Comulate maps a customer's full process (often 30-box flowcharts), then simplifies to 3-4 steps using AI.
• Simplification before automation is the step most competitors skip.

Why unfamiliar industries are the right place to look

• Industries historically untouched by software hold the largest pockets of manual work.
• ~70% of a recent YC batch was building AI agents for vertical industries — validation that the thesis is now mainstream.
• If a seed raise would have killed the idea early, field validation is the antidote.
• Go on a personal mission to find the truth — no one else's "altitude" on a problem substitutes for direct discovery.