01
Your business knows more than your software does.
The knowledge that runs a service business is already there. It’s in the people who know the job, the conversations that keep it moving, and the judgment they’ve built over years. A service manager knows what “ready” means in their shop and which exceptions need a closer look.
Some of that knowledge makes it into software. Much of it stays in conversations or with the person who handled the work. We believe AI products should make that knowledge easier for the whole team to use, while respecting the experience behind it.
02
Capability alone doesn’t run a business.
We ran into this ourselves when we started an AI project to find sales leads. Even with a developer background, setting it up took one of us a full day. A business owner should be able to get useful help without becoming an AI expert.
Easier setup is part of the work. A useful product also needs current information, an understanding of the team’s process, and clear limits on what it should do. It has to handle exceptions, make its work easy to check, and give people a way to correct mistakes. Those are responsibilities we have to build for as models improve.
03
A useful record of the work.
We start with a specific kind of work and the people doing it. Their updates can become a shared record: what changed, who handled it, what’s waiting, and what needs attention next. That record gives the team a basis for answering questions, coordinating handoffs, and deciding what to do.
Our aim is a detailed record of the work, without making the team spend its day maintaining one. Keeping it accurate means making updates easy to give, changes easy to trace, and errors easy to correct. The record has to be useful to the people relying on it.
04
Built for a working team.
Dealerships, plumbing shops, HVAC companies, and other service businesses need tools that fit a working day. A new product has to earn the time it takes to learn and the money it costs to run.
We aim to make capable software affordable to operate and easy for a working team to adopt. That means building around familiar ways of communicating, reducing duplicate recordkeeping, and helping people get started without a long rollout. We want the value to show up in their daily work: less chasing information, clearer handoffs, and more time for the job itself.
05
Starting with dealerships.
Our first product, Recon, keeps dealership reconditioning moving from the service team’s group text. It has been in daily use at a dealership since April 2026. Teams can give Recon vehicle updates, record an RO number, move a vehicle to its next stage, and ask questions from the record.
Recon gives us a concrete place to develop this approach, with real handoffs, parts delays, and changing priorities. We’re learning how to fit AI into those routines and keep the record useful as work moves between people.
We’re bringing that approach to the trades with a small group of design partners. Each kind of work has its own language and processes to understand. If that’s your business, we’d like to hear how the work gets done.