OFFICIAL SURVEY CONTEXT
AI Adoption in UK Businesses (BICS 2023–2026)
Central Statistical Principle: Adoption Rate ≠ Economic Contribution
Survey adoption rates measure the prevalence of business use, not monetary gross value added (GVA) or output. A company paying £20/month for an off-the-shelf generative assistant is classified as an "adopter" alongside an enterprise investing £100m in proprietary AI infrastructure. Adoption surveys cannot by themselves solve the National Accounts monetary disaggregation problem.
AI Adoption Trajectory among UK Businesses (2023 Q4 – 2026 Q2)
Percentage of businesses reporting active AI use in operations (BICS Survey)
Source: ONS Artificial intelligence in UK businesses (S2)UK AI Economic Measurement Lab
AI Adoption by Industry Sector (2026 Q2)
Percentage of businesses reporting active AI use vs intense production use
Source: ONS BICS (S2)UK AI Economic Measurement Lab
AI Adoption by Business Size (2026 Q2)
Adoption gradient across employment size bands
Source: ONS BICS (S2)UK AI Economic Measurement Lab
Why We Disaggregate Products Instead of SIC Industries
Standard industrial classifications (UK SIC) classify an enterprise based on its principal activity. A bank adopting AI remains in SIC 64; a car manufacturer using computer vision remains in SIC 29.
Therefore, measuring the AI economy requires product-level disaggregation (CPA) to separate AI services and goods produced from general industrial output.