AI implementation, the FDE way
The AI implementation playbook
Most AI rollouts measure the wrong thing and land in the wrong place: tokens burned, seats active, a chatbot bolted to the homepage. This is the other way. Put AI in the plumbing, govern it, and measure it against a number your CFO already tracks.
I do this work forward-deployed, which means inside the business rather than from a slide deck. I sit with a team, find where the week and the revenue actually leak, and put AI against those spots. Drafts, never sends. Then I stay until the team can run it without me.
The method below is free to read, start to finish. Four short playbooks: the stance, the governance-first setup, the way to hand AI your numbers, and an honest read on what today's AI can and cannot do yet. The implementation is the part I charge for.
The series
Common questions
What is forward-deployed AI implementation?
I work inside your business, not from a deck. I map where your week and your revenue actually leak, put AI against those spots, govern it, and stay until your team can run it without me.
How do you know whether the AI is working?
Against a number that shows up in the business: revenue, hours saved, faster turnaround. Token counts and active seats look busy without proving much, so they do not count as the result.
Do you build customer-facing chatbots?
Mostly no. The work I care about sits in the infrastructure: drafting, triage, research, the internal grind. AI that talks straight to your customers is a small and careful slice of it.
What does the AI Sprint include?
Six weeks: audit the workflow, deploy AI where it pays, govern it (drafts, never sends), train the team, and measure it against a real number. The four playbooks here are the method, free to read. The Sprint is the implementation.
The writing is free. The build is the AI Sprint.
If you want the method run inside your own workflow, that is the AI Sprint: six weeks to audit, deploy, govern, train, and measure. I take on a few at a time.