Artificial intelligence
AI adoption without the theatre
The gap between businesses that have an AI strategy and businesses that have AI in production is now the most expensive gap in the mid-market.
By Umar Ahmed, Founder, Nuub
·7 min read
Almost every leadership team we meet has been presented with an AI strategy in the past eighteen months. Very few have anything running in production that a finance director would recognise as a return.
The pattern is consistent. The strategy identifies twenty possible use cases, ranks them on a two-by-two, and recommends a centre of excellence. Nothing in that sequence produces a working tool, and the momentum dissipates within a quarter.
A more useful approach is narrow and evidential. Pick two use cases where the current process is measurable — time taken, error rate, cost per unit. Establish the baseline before you change anything. Implement into live use with the people who do the work. Measure the same numbers again at six weeks.
That approach produces something a board can act on: evidence. It also produces the internal confidence required for the next round, which no strategy document has ever achieved on its own.
The governance question is not optional. Any output that reaches a client, a patient or a financial record needs a named human accountable for it. Build that gate in at the start; retrofitting it after an incident is considerably more expensive.
Sources and evidence
Every figure in this piece comes from the evidence below. Each source was checked by a named editor before publication.
- 1.Nuub delivery practice: AI strategy against AI in productionNuub
In the engagements Nuub has run, adoption that starts from a measured baseline on two narrow use cases reaches live use more often than a portfolio-wide strategy does. Practitioner observation, not a measured study.
Nuub analysis·Unrated·Checked 20 August 2026