AI Transformation Field Guide
Redesign the work system. Not the task where AI enters.
Most AI programs bolt models onto existing tasks and hope value appears. This guide treats transformation as organizational design: the outcome-producing work system—roles, coordination, expertise, authority, and proof.
What this guide is
Practical. Opinionated. Built from deployments that worked and ones that didn’t.
Operator-written
Not a vendor deck or framework overview. Written by someone who has shipped AI into real work systems.
Work-system unit
The unit of transformation is the outcome-producing work system—not the task where the model enters.
Evidence discipline
Green dashboards don’t prove the system works. Evidence that survives examination does.
Free and complete
No gates, no drip, no fake unlock. The full guide is public. Use it to decide if you want to work together.
The problem
Task faster is not proof the system got better.
Teams ship copilots, recommendations, and agents into workflows that were never redesigned. Sellers keep the spreadsheet they trust. Underwriters keep the informal queue. Managers keep the approval ritual. Local metrics look fine. System outcomes barely move.
The useful question is not “Where can we use AI?” It is “How should this system of work change now that AI exists?”
That shift—from insertion to redesign—is what this field guide is for.
Chapters
Seven steps from understanding to operating.
Read in order when you can. Jump in when you know the failure mode.
01 · Start here
Understand the Work
Why most AI programs start at the wrong level—and how to see the outcome-producing work system before you redesign it.
Read chapter →
02
Decide What Should Change
From 'where can we use AI?' to a value mechanism, a binding constraint, and a real redesign choice.
Read chapter →
03
Redesign the Work
Roles, workflows, coordination, expertise, and incentives—so the system reinforces the outcome instead of fighting it.
Read chapter →
04
Design the Human–AI System
What software, models, and humans each do—and with what context, tools, permissions, and recovery.
Read chapter →
05
Prove It Works
From outcome to behavior to evaluation to evidence — so green local metrics cannot stand in for system proof.
Read chapter →
06
Govern It
The decision, not the system, is the unit of governance. Applicability, controls, evidence, accountability.
Read chapter →
07
Operate & Adapt
How the system stays honest after launch—monitoring, drift, workarounds, incidents, and adaptation.
Read chapter →
Practitioner artifact
Test the work before you build the AI.
Describe one piece of real work — the outcome, how it is actually done, what it quietly carries — and AI Work Fit returns an intervention call and then argues with it, using the hard case you wrote. No email, no gate.
Ready to work on a specific system?
The Field Guide is the method. If you have a work system that needs direct help, let’s talk.