Why I'm Betting on Vertical AI: Lessons from 7 Products

Everyone's building AI wrappers. Most of them will fail. Here's why I'm building vertical AI instead, and why I think it's the biggest opportunity in software right now.

Horizontal AI tools — chatbots, writing assistants, general-purpose automation — are in a brutal race to the bottom. OpenAI, Google, and Anthropic will eventually offer these capabilities natively. There's no moat in wrapping an API.

Vertical AI is different. When you build for a specific industry, you're not just wrapping an API — you're encoding domain expertise. AgriIntel doesn't just analyze data; it understands crop cycles, soil conditions, and market timing. CovioIQ doesn't just process text; it knows Canadian insurance regulations.

The moat in vertical AI comes from three places: domain-specific data, workflow integration, and trust. A farmer won't switch from AgriIntel to ChatGPT because ChatGPT doesn't understand their specific operation. An insurance broker won't use a generic AI tool because it doesn't know FSRA compliance requirements.

After building seven products across agriculture, insurance, real estate, compliance, and pet care, I've learned that the best vertical AI products start with a deep understanding of how an industry actually works — not just what it looks like from the outside.

My advice to AI founders: pick an industry, talk to practitioners, and build something that solves their specific problems. The opportunity is enormous and largely untapped.