Perspective · Thinking about AI adoption?
Before you spend more on AI
The pilots succeed and the organisation still does not change — that is the trap most adoption stories hit. The reasons are economic, not technical: where the cost actually sits, what it takes to cross from demo to value, and the judgement you stop manufacturing when you automate the work that used to teach it. This arc is the clear-eyed read.
A clear-eyed read on the economics of adopting AI — including the judgement you quietly spend down as you automate.
The pilots keep succeeding, and somehow the organisation keeps not changing.
It’s an economics problem, not a technology one — and automation quietly bills you in judgement, not just money.
Adopt on evidence, not enthusiasm: account for both halves of the P&L and the apprenticeship you’re automating away.
The Arc
7-step route
Begin → Why AI Pilots Stall Before ROIWhy AI Pilots Stall Before ROI
Upcoming · 10 minStart with the trap: pilots succeed, organisations fail — adoption is not value.
The Missing Rung
Upcoming · 9 minThe hidden cost of automating the first rung of the ladder.
Rebuilding the Apprenticeship
Upcoming · 11 minThe fix: design the apprenticeship the work used to provide for free — judgement was never free.
Build, Buy or Generate: How Agentic Delivery Rewrote the Oldest Capital-Allocation Decision
Upcoming · 8 minThe third path — generate — and the ownership liability it adds.
The AI Delivery P&L: Why Generation Got Cheap and Your Cost Base Did Not
Upcoming · 7 minWhy the cheapest part of the problem is the part boards budget for.
The Unit Economics of an AI Product
Upcoming · 11 minThe other half of the P&L: what it costs to RUN the product for every customer (inference is COGS).
AI Coding Governance That Enables, Not Forbids
Upcoming · 7 minAdopt on evidence, not enthusiasm.
Adopting AI well is an exercise in evidence and economics, not enthusiasm — and in noticing what you spend down as you automate. Governing those calls deliberately is the Decision Doctrine.
“Adoption is not value.”