There's a version of the AI-in-insurance story that everyone has already heard, and it ends with fewer people. We think that story misreads what claims work actually is, and we think the misreading is expensive, because it makes carriers optimize for the wrong scarcity.
The scarce resource in a claims department was never judgment or care. Adjusters have both in abundance; it's close to a personality requirement for the job. The scarce resource was time, and the time was being eaten by reading. A property claim arrives as hundreds of pages that all might matter: the report, the estimate with its dozens of line items, the policy and the amendments that change what the policy means. Someone has to get through all of it before a single good decision can be made. For decades, that someone was the same person who was supposed to be calling the family.
The trade nobody chose
Watch a good adjuster on the phone with a policyholder whose kitchen is under a tarp. That conversation (patient, precise, honest about what happens next) is the most valuable thing the insurance industry produces. It's the moment the abstract promise on the declarations page becomes a person keeping their word. And it's the first thing that gets cut when the queue is forty files deep, because the reading doesn't wait and the prompt-pay clock doesn't care.
Nobody chose that trade. Ask any adjuster what they'd do with a free hour on a claim, and the answer is never "read the amendments again." The documents chose it, because the documents were the only path to a defensible decision and nobody else could read them. That last clause is the one that just changed.
The design question we ask instead
So the question we ask about every feature is not "what does this automate?" but "what does this hand back?" A platform that has read every page hands back the phone call. A finding that carries its clause and its dollar impact hands back the confident explanation: the difference between "I'll have to look into it" and an answer with the policy language on the screen. A pre-flighted letter hands back the evening. The machine gets the reading; the person gets the parts that were always human, finally done with enough time to do them well.
This is also, quietly, a retention strategy. Ask why young adjusters leave the profession and the honest answers are drudgery and fear: drowning in documents, and dreading the miss that surfaces in a file review two years later. Remove the drowning and the blind spots, and the job that remains (investigation, judgment, helping people through the worst week of their year) is the job people thought they were signing up for. The capability case is the same thesis wearing a workforce hat.
AI won't make us less human. Used well, it can make us the most human we've ever been. We named the company's whole thesis after that sentence, we put it in the footer of every page, and we intend to be held to it by the claims teams we work with, and by the families on the other end of their phone lines.
