AI consultant. Licensed general contractor. I run both, every day.
Day to day I run open-weight models on hardware I own, build workflow automation the way n8n does it, run an ERP that holds the job costing and the client files, and point coding agents at my own machines to build the tools I actually use. A frontier model only sees a task when the work needs that level of reasoning, and even then it goes out through an API, on my own account, with only the minimum approved context for that task sent, never the client's full underlying dataset. I install every piece of this myself, and I am the one who maintains it.
I run a construction company, an electrical company, and a cultivation and packaging operation. Every job in each of them runs on numbers: the cost book, the bids, the client files, the schedule. For years those numbers lived in software I paid a monthly fee for, on a server I never saw, run by a company that could raise the price whenever it wanted.
So I built a private AI system for my own companies first. A computer in my own building, running open-weight models, doing the same work the cloud tools did, without handing my day-to-day data to somebody else's server.
It worked well enough that other business owners started asking how I did it. A clinic owner. A staffing agency. A law firm. Each one had the same problem in a different coat: somebody else holding their client data and charging them monthly for the privilege.
That is the moment this became a second business. Not a software company. An operator who installs the same system he runs, in your building, and stays on to watch it.
What I am not
I am not an agency. There is no account manager. You get me.
Who this is for
Trades and home services get their own page too.