For one of our team offsites, we flew the whole team from New York to Germany for steins, lederhosen, and a proper beer-festival weekend. Between toasts, we took a customer call at 11pm here and there (California doesn't run on Oktoberfest hours). That photo is the short version of who we are: a small team that mostly feels like a group of friends, working intensely on a problem we're obsessed with, and having a lot of fun doing it. We're an AI company for one of the most traditional industries on earth, and we use AI for nearly everything else. If "I'll have my Claude reach out to your Claude" reads as a normal sentence to you, you'll fit right in.
Oktoberfest team offsite 🇩🇪
We build AI agents for the back office of insurance claims: the phone calls, emails, and documents adjusters spend their days on. We ended up here because it's one of the most overlooked corners of the economy. More than a million people do this work, the software they've been handed is awful, and a lot of sharp people are stuck with bad tooling. It's a huge market everyone else found too boring to fix. Insurance is nearly twice the size of the whole tech industry: about $9 trillion in revenue a year. The work is repetitive and structured, which makes it a near-perfect fit for AI, and we're early.
We closed our first customer with a cold walk-in. During YC, one of us packed a suitcase, flew across the country, and walked into their office to ask one question: do you have any problems we could solve with AI? They did, and have been an incredible customer since.
We still work that way. Insurance runs on relationships, and relationships happen in person, so we show up: at the conferences, in the customer's office, at the bar afterward, on the golf course the next morning. From the outside the industry looks stuffy and old-school. From the inside it's full of sharp, funny, warm people, and we're usually the youngest in the room. Some of our best nights as a company have been karaoke at 3am with insurance executives. That's the energy we want to keep as we grow.
Getting our first customer via cold walk-in (during YC)
Day to day, we get excited easily: a customer going live, a new model dropping, a clean JSON schema, a parser that finally reads a document three people had given up on. We try things before we judge them, since most of what gets said about insurance and AI is wrong in both directions. A lot of us are German, which shows up as directness: we don't dance around things, feedback is fast, you'll always know where you stand, and we'll take a plain answer over a polished one.
We're in the office five to six days a week. The work is intense, the hours are flexible, and we're serious about having fun while we do it: lunch by the river, drinks after work, volleyball or a run when someone calls it. Life is short. If the work is going to be this intense, it had better be fun too. It isn't for everyone, and we'd rather say so now.
We don't chase the perfect solution. We ship the smallest version that works, then iterate really fast, until the problem is solved.
Crisp doc, messy meeting, decisive when it counts. We write things down to think them through, argue them out in the room, then someone calls it and we move. We run retros often and change our minds in public when the evidence says to.
Ownership comes fast. One of our forward deployed engineers joined on a Monday while we were sprinting on a customer POV. Before we'd even done the normal onboarding, we pulled him onto the project. He got up to speed in a week, demoed to the customer's CEO six days in, then flew to the middle of the country with us. We signed a six-figure partnership that week. He still owns that relationship and the implementation today. Everything here is up for grabs: responsibility, growth, the chance to run your own thing.
Our core principles are:
We've grown fast since YC, the seed is in the bank from top VCs (Frontline, 1984, L2, Booom), and a Series A is on the horizon.