What I've built
A personal AI operating system
A context layer for AI agents. It holds what I know and what I'm working on as a linked knowledge graph in markdown, so every session starts with context instead of a blank page. An agent reads it, keeps it current and drafts my next steps, with a human in the loop on anything that goes out. Built with rules loaded on demand and an audit log of every action. A team version is in pilot at work, served to teammates' assistants over MCP with vector search.
An AI assistant for 1,500 field reps
Answers, deal guidance, and a heads up when a deal stalls, all in Slack and connected to the systems that hold the data. We tested it against a set of real questions, which showed the limit was the knowledge base, not the model.
A lead-to-revenue pipeline
Ads, call tracking, a call center, and a field CRM connected into one path from first click to booked revenue, offline sales included.
A searchable API reference
A reference for an API with thin docs, where you can try calls right in the browser. AI agents can read it too, and a community of operators now uses it as their shared reference.
Identity automation
Accounts that open and close on their own when someone joins or leaves, across four systems that didn't agree. Lining them up also found people each system had lost track of.
A monthly builders meetup
A monthly call I started and host, for operators and developers building on the same field service API. It's become the group's regular place to share tools.