Learning Hub
What a stalled AI rollout actually costs, in plain pounds
I take one typical £50m business and run it through all four ways a rollout goes, with the pounds worked out in full. Straight answers for the CEO who has to face the board in the morning.
Case studies
One £50m business. The same £1.5m AI bet. Four very different endings.
I built one reference business and ran it through the four friction states, with the maths worked out in full each time. Three quietly drain it. One pays for itself.
- The Stewardship Dividend +£2.5mEBITDA Productive Friction Aligned and free to push back. That’s Productive Friction. The rollout goes live in 90 days, frees £1m of capacity, and lifts the top line to £55m. The only ending that pays. Read the case study
- All gas, no speed −£460kEBITDA The Deceleration Tax Always nearly ready, never live. A faster rival ships first and poaches £3m of revenue while the firm runs one more readiness workshop. Read the case study
- Unregulated Adoption −£630kprofit hit The Shadow AI Tax Your people use AI on the quiet, no protocol, no oversight, no alignment. 49% of workers admit it, and 51% have wired unmonitored tools straight into work systems. Every link is a liability. Read the case study
- Playing it safe −£200ka year The Hedging Tax Everyone agrees, nobody adopts. The £2m boost never arrives, the spend lands anyway, and dual-running becomes a permanent drag. Read the case study
From the blog
Notes on why rollouts stall, and what it’s costing
A new and growing set. Search the titles or filter by topic to find the question keeping you up.
1 article
No articles match that yet. Clear the search and try again.



