Maison — AI Strategy Advisory
AI counsel, written by the hand that builds it.
Most technical decisions are handled by the team you have. The few that are not, such as strategic platform bets, vendor decisions, and the direction of your operations, are the ones where a wrong move can cost three years and a right one compounds. Those are the decisions where counsel earns its keep.
I've founded four companies, most recently Kacti AI, where I lead applied AI R&D, and spent fifteen years at Microsoft, most recently as a Principal Architect on the Dynamics 365 VP/CTO team. I take on a small number of engagements each month. If the timing is right, we should talk.
Point of View
Three convictions that shape every engagement.
— 01
The hard part of an AI program is rarely the engineering. It is the selection.
By the time someone is debugging a prompt, the consequential decision has already been made — what to build, in what order, at what scope. Failure happens at the moment of choice, not the moment of build. The work that matters lives there.
— 02
The quality ceiling of most organizations is not set by capability. It is set by accumulation.
Every team inherits processes, tools, and integrations from the organization it used to be. Each one made sense at the time. Together they form a ceiling — a cap on how correct, how complete, and how valuable any decision downstream can be. The highest-leverage work is rarely addition. It is subtraction.
— 03
The biggest risk in enterprise AI is not hallucination. It is silent incompleteness.
These systems retrieve fragments and present them as complete answers. Your team makes decisions on partial information that looks whole, with no mechanism to flag what was missed. The cost surfaces later — in a forecast that drifts, a board report that doesn't match the field, a quiet loss of trust no one can trace to a single error.
24+
Years in Software R&D
Named
Inventor, enterprise software patents
4
Companies Founded
15
Years at Microsoft
Begin
Worth a conversation.
If you're facing a strategic AI or technology decision that will impact your organization for the next three years, fifteen minutes will tell us both whether there's a fit.