UK enterprise AI adoption doubled in a year, from 29% to 60%, but a new SAS and IDC survey points to a gap between deploying AI and getting value from it. Its practical message for organisations: measure what AI changes in a decision, and match how much freedom a system gets to the quality of the data and controls behind it.
Watch Desk analysis
What happened
The second edition of the SAS and IDC Data and AI Impact Report surveyed 2,699 decision-makers across 28 countries. The report puts the global Trustworthiness Index at 61.3, up 11.6 points, and says organisations with trustworthy AI practices were 15 times more likely to report strong or high returns on investment than those without them: 62% compared with 4%.
AI Magazine interviewed Dr Iain Brown, SAS’s global head of AI and data science, about the findings. For the UK, the report gives an AI Impact Index of 51.7, 8.8 points below its Trustworthiness Index. Data quality and governance scored 56.5, the UK’s lowest result across the trust dimensions. These are survey findings and reported associations, not proof that trust practices alone cause higher returns.
Why it matters
Counting launches is easier than showing that an AI system improved a decision, reduced friction or created measurable value. Brown argues that organisations should work backwards from the decision an AI system will influence, then identify who owns the outcome and whether the supporting data is suitable, traceable and monitored after deployment.
His proposed middle course is progressive autonomy: use AI now, but keep tighter boundaries and human approval where data quality, lineage or explainability are weak. Grant more freedom only as evidence and controls improve. That distinction matters especially for agents that can take actions, not merely offer recommendations.
Our read
The useful advice here is not to wait for immaculate data. It is to stop treating deployment as the finish line, and to make the system’s authority answer to the evidence beneath it. A chart about trust is not a substitute for an outcome measure, but it is a nudge towards asking better questions before the pilot becomes everyone’s workflow.
What to watch
- Whether organisations measure changes to decisions and outcomes, rather than counting launches.
- How they assign ownership for data quality and AI-influenced decisions.
- Whether higher autonomy is matched by stronger validation, explainability and controls.
Discussion spark: Should companies let AI systems act before their data and controls are mature, provided people approve consequential decisions, or keep those systems advisory until the foundations improve?
Sources and evidence
- Iain Brown, SAS AI Head: Turning Deployment into Impact – AI Magazine (6 October 2026, 09:36 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.