Find the leverage
I map how the work actually happens, then pinpoint the workflows where AI removes real hours instead of adding novelty — and the ones where it shouldn’t be anywhere near.
In the org
I help teams put AI where it actually changes the work.
Most AI initiatives stall at the demo. I close the gap between “impressive” and “in production” — finding the high-leverage workflows, building the tooling, and bringing the team along so the change sticks after I leave.
What I do
Three ways I move AI adoption from slideware to shipped.
Build the tooling
Prompts, assistants, MCP servers, internal apps, and pipelines — built end to end and wired into the systems people already use, so the win shows up in the daily workflow, not a sandbox.
Enable the team
Hands-on training, patterns, and guardrails that turn “the AI person” into an org-wide capability. The goal is a team that keeps shipping with AI without me in the room.
How I operate
A bias for shipping, measured against the work — not the hype.
-
Start from the workflow, not the model
Shadow the real process first. The question is never “where can we use AI?” but “where is the work slow, error-prone, or repetitive enough to be worth it?”
-
Ship a thin slice fast
Get something real in front of real users in the first weeks. A working slice beats a strategy deck for learning what actually helps.
-
Instrument and measure
Define the before/after metric up front — time, throughput, quality, cost — and track it, so adoption rests on evidence rather than vibes.
-
Transfer ownership
Document the patterns, train the team, and set the guardrails. Success is measured by what keeps running once I’m gone.