Insurance is all text. Contracts, regulatory filings, processing every single claim. Exactly the kind of thing LLMs are best at. Corgi's founder says as much: the product found product-market fit so fast precisely because of the industry's "text-based nature."
How Corgi works
In two years, Corgi became a unicorn valued at $1.3B, insuring over 40,000 clients across 49 states. This isn't an app slapped on top of someone else's carrier — it's the first AI-native licensed insurer in the US. They own the full stack: insurer, reinsurer, and platform in one. No middlemen between you and your policy.
They insure startups against their most common headaches. If investors or employees sue founders over management decisions, the policy pays out — not the founder's personal savings. VCs almost always require this coverage; you can't close a round without it. Cyberattacks and customer complaints about how the tech performed are covered too. But the more interesting part is coverage for AI risks themselves. Legacy insurers are busy excluding anything AI-related from their policies right now — Corgi is built on AI and is betting on insuring it.
The company is about 100 people. Where a traditional insurer has thousands of people manually reading contracts and calculating risk, Corgi has a model doing it. A policy gets assembled in a few minutes; coverage options flip on like feature toggles in a software dashboard.
Growth strategy
Startups are the entry point, not the whole market. Corgi is already porting the same engine into cargo insurance, with payments and small business next. Once you've rewritten an insurance company as software, moving into a new vertical is just changing the settings. Legacy players can't do that — each market has its own legacy systems and its own people in suits.
Which is a pretty good hint for what to automate next. Look for industries that seem boring and calcified from the outside, but on the inside run entirely on text-based busywork and manual document review.
The same pattern in other industries
Vertical plays like this already exist:
- In law, [Harvey raised at an $11B valuation in March](https://www.cnbc.com/2026/03/25/legal-ai-startup-harvey-raises-200-million-at-11-billion-valuation.html): same contracts, due diligence, and court filings handed off to a model.
- In medicine, [Abridge at $5.3B](https://techcrunch.com/2025/06/24/in-just-4-months-ai-medical-scribe-abridge-doubles-valuation-to-5-3b/) turns doctor-patient conversations into ready-to-file clinical notes.
Same pattern everywhere: a mountain of text that used to take forever to process by hand.
*The Corgi team works seven days a week, and new hires get a mattress after a month. Further confirmation that with AI, we're all just going to be working more.*