Key findings
What the evidence shows
- 35%
- of UK businesses with ten or more employees now use AI, up from 12% in 2023.
- Office for National Statistics, July 2026
- 1 in 10
- businesses that use AI describe their use as extensive.
- Office for National Statistics, July 2026
- 89%
- of firms in a four-country survey of nearly 6,000 executives report no effect on productivity over the previous three years.
- NBER Working Paper 34836, February 2026
- £428,000
- is the illustrative annual value of a 4% productivity gain to a business of 50 people.
- The Thirty Co. calculation from BIS and DBT data
Contents
Inside the report
Section 1
The lever
The UK productivity gap in official figures: a small group of leading firms and a long tail behind them.
Section 2
Adoption is not implementation
What the ONS, DSIT and an international survey of nearly 6,000 executives found beneath the headline adoption figures.
Section 3
What effective implementation looks like
Five disciplines applied in sequence, with a first step for each.
Section 4
The evidence
Our analysis of more than 750 communications between business owners, set against published field research.
Section 5
The national opportunity
What IMF staff modelling and the OBR scenarios suggest about the scale of the gain.
Section 6
What to do
The support available, three steps to take, and five shifts in practice.
Section 7
What the gap is worth to one business
Published research and official statistics, expressed in pounds for a single firm.
Section 8
Common objections
Five objections to implementing AI, and what the evidence says about each.
Who the report is for
The report is written for the person accountable for how a business performs, whether you own it, run it or run a part of it. It is most relevant where AI tools have been paid for and are in use, and the effect is not yet visible in the accounts.
The evidence base
The report draws on published work from the Office for National Statistics, the Department for Science, Innovation and Technology, the International Monetary Fund, the Office for Budget Responsibility and the Bank for International Settlements, together with our own analysis of more than 750 communications between business owners using AI. Every source is listed in full.
From the report
Related insights
Analysis 8 October 2026
Adoption is not implementation
The number of UK businesses using AI has almost tripled. One in ten of them describes its use as extensive. The distinction explains why most firms are yet to see a return.
Read moreGuidance 8 October 2026
Three common mistakes in AI adoption
Most firms report no effect on productivity from AI so far. The evidence points to three recurring errors, each of which is a reasonable decision when taken on its own.
Read moreAnalysis 8 October 2026
What the implementation gap is worth to a business of 50 people
National estimates of the value of AI are difficult to relate to a single business. Using published research and official statistics, the figure can be brought down to that scale.
Read moreGuidance 8 October 2026
Five objections to implementing AI, and what the evidence says
Five objections are raised more often than any others. Each has a basis in fact, and each has an answer in the evidence.
Read more

