One owner, three shapes of engagement — chosen by what will actually create value inside your organization, not by which deliverable sells easiest.
When the problem is messy and there's no clean spec yet.
I start most engagements here. I read the workflow, talk to the people who live with it, and come back with a map of where AI should act on its own, where humans stay in the room, and why each line sits where it does. I ground it in Information Systems research on cost, efficiency, and adoption.
When the problem is messy and there's no clean spec yet.
I own architecture and implementation end-to-end. I build the agents and pipelines where automation pays, the interfaces where human judgment does, and the joints between them — no handoff chain between the person who understood the problem and the person writing the code.
When the target is defined and you need it shipped, not re-scoped.
I stay with the system past the demo. I measure adoption, fix the joints that don't hold under real load, and evaluate against the goals we set at the start. I apply research discipline to production — so the system keeps earning its place six months in.
When your last AI pilot didn't survive contact with real users.
I bring the research into the room. I run half-day to multi-day formats on user-centered AI, knowledge graphs, and how to decide what to automate and what to keep close to human judgment. I write them for product, research, and operations people who want the lens, not just the deliverable.
When you want to invest in your own judgment.
When the target is defined and you need it shipped, not re-scoped.
Contact