What I'm working on right now.
Building AI systems for companies that are ready to stop experimenting and start integrating.
The pattern across every engagement is the same: find where AI creates compounding leverage, build it fast, make it stick.
Client engagements. Building AI-powered operational tools for mid-market companies. The details are confidential, but the patterns are worth noting: most companies overestimate the technology challenge and underestimate the integration challenge.
Software of you. My own personal OS: CRM, email triage, task management, knowledge base — all AI-powered, all local-first, all running daily. I use it to manage every client relationship and every piece of work. If I'm going to tell you to integrate AI into your operations, I should be doing it myself.
Writing. Working through ideas about the integration-vs-adoption distinction, what mid-market companies get wrong about AI, and why the next 12-18 months matter more than the last 12.
- Why “AI strategy” is usually a delay mechanism, not an acceleration one
- The difference between AI tools that are nice-to-have and ones that are load-bearing
- Multi-agent orchestration patterns — how multiple AI systems coordinate on complex tasks
- What happens to companies that wait