An endurance race · Eight episodes
What happens to a delivery organization when AI makes one stage of the pipeline faster and nothing else changes.
Le Mans runs for twenty-four hours. Hour eighteen lands the next morning, and the night — the cold, the traffic, the hour when things break — is behind you. Nobody in the garage has slept. The sun is low enough to be a problem and the screens are washed out in the glare. And eighteen hours of watching a car go round has made even the people paid to watch it stop really looking.
It is where the race is decided — and it is decided in both directions. The cars that stop, stop around here. So do the ones that go on to win.
Map that onto a four-year technology adoption and hour eighteen is year three, which is roughly when the consequences of the first year's decisions become measurable.
Enterprise AI adoption doesn't fail in the build phase. It fails at hour eighteen — in the functions nobody upgraded.
This series takes the build phase as the starting point and walks outward in both directions at once, because that is how the pressure actually propagates. Each episode is a short read.
Finding a way to winAudi won Le Mans thirteen times in fifteen years. Not by building a car that would not break — at racing load nothing survives twenty-four hours untouched — but by building one that could be repaired faster than anyone else's, and by deciding in advance where it could afford to fail.
The failures described here are real and mostly unbudgeted. They are also known, locatable and survivable, which is the difference between a warning and a plan. Each episode closes with the line through that corner; from Episode 5 the series is about what to do rather than what goes wrong.