Read the insurance AI conversation on any given day and you'll notice something: almost everyone talking is selling something. Vendors selling platforms. Consultants selling transformations. Analysts selling reports about the vendors and consultants.
Almost nobody is writing from the operator's chair — the seat where the budget, the systems, and the outcome all belong to the same person. I sat in that chair as a carrier CIO. I know what that conversation sounds like from the inside, and it sounds nothing like the vendor version.
So this is the counterweight. One short note a day: what happened in insurance AI, what it actually means for carriers, and what I'd do about it. Operating models, governance, economics — the three things that decide whether AI survives contact with production.
Almost nobody is writing from the operator's chair — where the budget, the systems, and the outcome all belong to the same person.
I'm also writing a book — The Insurance AI Operating System — and these essays are its first draft, written in public. Some of them will become chapters. Most of them won't. All of them will be honest.
Three rules for this notebook: no vendor pitches, no hype cycles, no invented numbers. If I don't know something, I'll say so. If I'm wrong about something, I'll say that too, the next day, in writing.
— Krishna