PARTICIPATION
Anonymous by default
The benchmark does not require an organisation name or personal contact details for the research response.
Raeburn Institute · Open research programme
UK Enterprise AI Governance Benchmark
A longitudinal benchmark of how UK organisations govern AI in practice. Organisations are invited to complete an anonymous assessment; eligible responses are analysed only in aggregate and the methodology, limitations and annual findings are published openly.
REPORTING
Aggregated findings
Results are reported at cohort level. Small segments are suppressed where disclosure risk would be unreasonable.
CADENCE
Annual publication
A dated annual report will be published when the response base is sufficient to support meaningful analysis.
Benchmark dimensions
1.AI inventory & ownership
2.Risk classification
3.Data governance & privacy
4.Evaluation & assurance
5.Human oversight
6.Agent/tool controls
7.Third-party/model governance
8.Change management
9.Incident response
10.Value & lifecycle governance
What participants receive
Participants can calculate their own governance maturity score immediately. Once a sufficiently robust benchmark cohort exists, the programme can publish comparative distributions by organisation size and other broad non-identifying cohorts. Until then, no claim of UK representativeness or percentile ranking is made.
Research integrity
This is a voluntary-sample benchmark, not a census and not certification. Annual publications will state sample size, fieldwork dates, cohort definitions, exclusions, missing-data handling and material limitations. Raw organisation-level responses will not be published as part of the public benchmark.