Archives
All the articles I've archived.
-
The Unglamorous Work
The work nobody talked about as 'AI work' — underwriting queues, regulatory scanning, SKU reconciliation — is where AI is paying back fastest. Three cases and a pattern.
-
The Forcing Function
Internal roadmaps rarely drive production AI deployment. External forcing functions — a regulation, a volume shock, an embedded tool — do. Three cases that show what it actually takes.
-
The Constraint Was Never the Obvious One
The most consequential AI deployments share a pattern: they didn't apply AI to the existing process. They first identified what the existing process was actually built around — and attacked that.
-
The AI Transformation Nobody Noticed
Five energy companies — Ørsted, Shell, BP, Enel, Xcel — quietly redesigned how they operated around what AI made possible. One pattern underneath all of it: the decision always came first.
-
The Lunar Winter: What It Took to Go Back
NASA had the capability to return to the Moon for fifty years before anyone did. The gap wasn't technology — it was the conditions that turn capability into action. Every organization has its own lunar winter.
-
Stop Automating the Wrong Thing: A 50-Year Pattern That Explains Why Most AI Deployments Disappoint
Ford did not speed up invoice processing. FedEx did not make deliveries faster. Progressive did not schedule adjusters more efficiently. Each one noticed a technology had opened a new door — and walked through it. AI is opening those doors right now.
-
The Data Moat Playbook: How Non-Tech Incumbents Won the Internet Era — And the Lesson for the AI Era
How non-tech incumbents built moats nobody noticed — and why it's happening again right now