I'm a fractional COO. I run operations for a handful of companies at once.
My own work runs on an agentic fleet I built. That's how I run more than one operation without the work waiting on me. I've run companies across pharma, devices, and software for twenty years; over the last two years I built the fleet and the practice around it.
The leverage goes to the people who know the operation. When nobody in the room can evaluate what AI built, the work breaks there.
- Standing agents 20–50 workspaces
- Durable context memory · repos · task lists
- Running services backends · timers · triggers
The fleet stays running. I step in where the work needs judgment.
Why now
AI is in a honeymoon phase. One camp thinks anyone can build an enterprise application in an hour. Another thinks it will replace the experts.