How are UK businesses scaling AI beyond pilots?

Almost every UK organisation now has AI running somewhere. Very few have scaled it. In aibl’s survey of 755 UK mid-market business leaders, 87% expect AI to transform their industry and 82.5% have a board mandate to invest in 2026. Yet only 14% have deployed AI at scale, and about half can show a measurable return.

What separates the businesses that scale AI from the rest is not budget. It is a named owner, governance that is followed, and a baseline set before each deployment.

Why do most AI programmes stall before they scale?

Only 14% of UK mid-market organisations have deployed AI at scale, even though 65% can approve a new AI tool in days or weeks. Speed of buying is not the constraint. Three things hold programmes back:

• No single owner. Across the survey, AI with no single owner returns just 18%
• A policy on paper that nobody follows, which delivers almost no improvement in return
• No baseline, so improvements cannot be shown as a return the finance team accepts

Should you go broad or deep first?

Go broad first. In aibl’s survey of 755 UK mid-market business leaders, organisations that spread AI across several functions report 57% measurable ROI, against 49% for those that went deep in just one. Map yourself on two questions, how many functions use AI and whether any has reached scale, and you land in one of four groups:

• Stalled: AI in one or two places, nothing scaled. 21% measurable ROI, and where most organisations sit.
• Siloed: deep in one function, narrow everywhere else. 49%.
• Scattered: AI across three or more functions, no deep scale yet. 57%.
• Scaled: breadth plus depth. 62%, rising to 72% once two or more functions are scaled.

Depth pays once the breadth is there. Basic tools and single-task agents return 29% measurable ROI, agents running a full workflow 56%, and agents working across departments 83%. So the sequence is breadth first, then depth: get a second and third function using AI properly, then push the strongest towards agents that run a whole workflow.

How does governance affect AI at scale?

Governance is the strongest predictor of AI returns in the survey. Only about one in five organisations has mature, embedded governance. Where governance is weakest, shadow AI runs at 75%. Where it is mature, it falls to 35%.

The step that matters most is moving from a written policy to an enforced one. Leaders with a policy that is not consistently applied report returns close to those with no policy at all. Our piece on how governance drives AI ROI sets out the full ladder.

Which functions are furthest ahead with AI?

Deployment and proof vary widely by function:

• HR and people: 91% have AI deployed, the broadest of any function
• Finance and operations: 87% have AI in place, with the best-governed teams proving the most
• Customer service and CX: 81% have AI embedded, with the highest failure rate
• Sales and marketing: the most momentum of any function, but only 45% can prove a return
• Technology and IT: 58% measurable ROI, the highest of any function

For a transformation leader, the gaps between functions are the programme. Bring every function up to the governance standard of the best one.

How should you measure an AI transformation?

Start with the AI already running. Pick the three busiest AI workflows across the business, define what success looks like in each and set a baseline. Then report against it every quarter to the executive team.

For every new deployment, refuse to sign the purchase until the success metric and the baseline are in the business case. Our guide to building an AI business case your CFO will approve shows how.

Where can transformation leaders get help with AI?

Partners listed under AI Strategy and Readiness in the aibl AI Enablement Directory help set priorities. Those under AI Implementation and Deployment take pilots into production, and AI Training and Capability partners build the skills to run AI day to day. Our guide on how to choose an AI implementation partner covers the questions to ask.

Go deeper: CEOs and boards have their own cut of the data. aiblLIVE brings transformation leaders and their peers together to compare what is working.

What this means for the transformation lead

Name one executive owner for AI outcomes. Check that your AI policy is followed in every function, not just written down. Then set a baseline for the three busiest AI workflows and report against it each quarter, so the next stage of investment rests on proof.

Frequently asked questions

Read the full research

This is the AI and transformation cut of State of UK AI Adoption 2026, aibl’s benchmark of 755 UK mid-market business leaders, in partnership with Executive Summary. The full report covers every function, the governance ladder behind these numbers, and the four moves that separate the firms proving a return from the rest.

Read the full State of UK AI Adoption 2026 report →

If you are leading an AI programme across the business, the AI Enablement Directory lists UK partners by category, including AI strategy and readiness, implementation and deployment, and training and capability. New findings land each Friday in aiblBRIEF, our free weekly newsletter.

About the research: the State of UK AI Adoption Survey 2026 was run with aibl’s research partner Executive Summary. It covers 755 UK mid-market business leaders in organisations with revenues of £50m to £500m, fielded January to March 2026. Every figure above describes the leaders surveyed.

For more on how UK businesses are scaling AI, explore our AI and transformation insights.