Who should own AI in your business?
The CEO or C-suite should own it. In aibl's survey of 755 UK mid-market leaders, companies with an executive...
Read moreYes. It’s the strongest predictor of return aibl found. In aibl’s survey of 755 UK mid-market leaders, 22% with no governance could show a measurable return, against 85% of the most mature. Same tools, same market. The difference is governance.
Governance is the single biggest factor separating UK businesses that can prove AI works from those that can’t. It isn’t about better technology. It’s about whether anyone owns the outcome and checks the number. In aibl’s survey of 755 UK mid-market leaders, measurable return runs from 22% with no governance to 85% at full maturity.
Line up the five governance levels and the pattern is stark. No governance sits at 22%. Informal guidance reaches 28%. A defined-but-inconsistent framework hits 32%. Formal governance climbs to 59%. Mature, embedded governance with monitoring and board review reaches 85%.
That’s a 63-point gap between the bottom and top rungs, on much the same tools throughout. Only 21% of UK businesses have reached that top rung. Most sit somewhere in the middle, doing some of the work and getting some of the credit.
The reality is that governance isn’t paperwork. It’s three things: someone owns the outcome, the success metric gets written down before the spend, and the number survives scrutiny from the CFO. Do all three and the return follows. Skip any one and it usually doesn’t.
Enthusiasm alone doesn’t move the number, and the data is blunt about it. Companies that are all-in on AI but ungoverned report just 39% measurable ROI. Companies that are well governed but far less enthusiastic report 54%. Governed and genuinely keen together reach 75%.
About a third of UK businesses sit in the ungoverned-but-all-in group. It’s the biggest single position in the market, and also the weakest. They’ve bought the tools and made the announcement. They haven’t built the structure to prove any of it worked.
That’s a hard pattern to sit inside. You feel busy. You feel committed. But without governance behind the commitment, the number the board wants never quite arrives. Frankly, appetite without accountability is a spending habit, not a strategy.
A written policy nobody follows can be worse than having no policy at all, and HR shows it most clearly. In aibl’s survey of 755 UK mid-market leaders, HR functions with defined-but-inconsistent governance report just 16% measurable ROI, against 33% for HR teams with no governance whatsoever.
Something similar shows up in technology and IT, where informal guidance (40%) beats a defined-but-inconsistent framework (33%). The middle rung isn’t a safe halfway house. It’s often where confidence outruns control.
The likely cause is false comfort. A framework on the intranet feels like the job’s done, so nobody spends the money or the attention needed to actually enforce it. Read our foundation guide on what AI governance means if your policy exists but nothing has changed since you wrote it.
Governance is one lever. Capability, meaning skills and confidence to use AI well, is the other. Governed and well-equipped companies reach 81% measurable ROI. Pull only one lever and you land around 40 to 42%. Pull neither and you’re at 15%.
That’s a meaningful finding for UK businesses weighing where to spend next. Governance without trained people stalls. Trained people without governance can’t prove their work paid off. You need both moving together.
This also explains why shadow AI, unapproved tools your team uses anyway, falls as governance matures. It runs at 75% in ungoverned companies and drops to 35% in the most mature ones. When the approved route works, people stop routing around it.
Good governance isn’t a document. It’s a named owner, a written baseline, and a fast approval route people actually use. Our AI governance framework sets out the structure step by step, from ownership to monitoring.
Start smaller if you need to. Put a name against who owns AI outcomes at executive level. Before the next tool gets signed off, write the success metric and where that metric stands today. That single habit, more than any policy document, is what separates the 22% from the 85%.
Governance done properly also changes how confidently your business can move. Two-thirds of leaders in aibl’s survey say they can approve a new AI tool in days or weeks rather than months, and faster approval tracks with higher measurable returns. Speed and control aren’t opposites here. They reinforce each other.
None of this requires a large team or a consultant-built framework. It requires the discipline to write the metric down before the money moves, and someone whose job it is to check whether the number showed up.
Yes, and the effect is large. In aibl’s survey of 755 UK mid-market leaders, measurable ROI rises from 22% with no governance to 85% with mature, embedded governance. That’s a 63-point gap on broadly the same tools, which makes governance the strongest single predictor of return in the data.
Because the tools are largely similar across UK businesses, but the discipline around them isn’t. Governance means someone owns the outcome, the success metric is written before spend, and the number gets checked. Without that structure, even good tools produce results nobody can prove or defend.
A named owner at executive level, a written success metric and baseline before any purchase, and a fast approval process people actually follow. Monitoring and board review mark the most mature stage. A policy that sits unused on an intranet doesn’t count, and can perform worse than no policy at all.
Start with ownership and one metric. Name who is accountable for AI outcomes, then pick a single use case and write down the metric that would prove it worked before anyone signs the purchase. That habit alone moves UK businesses further than a lengthy policy document ever does.
This is one finding from State of UK AI Adoption 2026, aibl’s benchmark of 755 UK mid-market leaders, in partnership with Executive Summary. The full report breaks down the governance ladder by function, ownership model and sector, with the moves that close the gap.
The CEO or C-suite should own it. In aibl's survey of 755 UK mid-market leaders, companies with an executive...
Read more
Lead with the number your CFO already trusts, not the tool a vendor is selling. Write the success metric and the...
Read more
Readiness for AI is not about the tools you buy. It comes down to two things the survey shows decide the return:...
Read moreGet ahead with the most actionable insights, playbooks and real-world AI use cases you can adopt right now, in your inbox every week