Every UK boardroom seems to agree that AI is transformative — the disagreement is over what "transformation" has actually delivered so far. The latest McKinsey State of AI survey reveals that 88% of all organisations around the world now employ AI in some part of their business, but just 39% claim to have seen positive EBIT impact — largely flat on the year before.
For many UK finance directors, that difference is one of the reasons they are now taking a more serious look at custom AI solutions UK teams build and own outright, finding them more believable than the one-size-fits-all tools they've all been racing to adopt first.
This report collates the key data points that are crucial for businesses in the UK to consider as they determine their next steps: the extent of adoption, where the ROI is (and isn't) coming from, how agentic AI is moving from pilot to production, and governance gaps holding back the transformation.
The State of AI Transformation in the UK, 2026
- The number of active AI businesses in the UK rose to 5862 (an increase of 58% year-on-year) with turnover of £23.9 billion and employment of 86,139 (DSIT AI Sector Study).
- 89% of organisations globally say they are making regular use of AI in one or more functions, with 44% now at enterprise-scale compared with 38% the previous year (McKinsey).
- The use of AI tools by workers grew by 50% in 2025 alone, according to Deloitte's State of AI in the Enterprise 2026 report which surveyed 3,235 senior leaders across 24 countries.
- 75% of financial services firms have adopted AI in the UK compared to 58% in 2022, and another 10% are set to adopt in the next three years (Bank of England / FCA).
- The UK government's own AI Opportunities Action Plan aims to deliver an extra £400bn of economic growth by 2030, supported by a 20-fold increase in the UK's AI Research Resource.
Where the ROI Actually Shows Up
This is the section most transformation reports skate over, and it's the one that determines next year's budget. Despite near-universal adoption, McKinsey's data shows only 37% of organisations attribute any positive EBIT impact to AI, and just 6% qualify as "high performers" — those seeing a 5%+ EBIT contribution with a clearly redesigned workflow behind it. Deloitte's figures tell a similar story from a different angle: 66% of leaders report productivity or efficiency gains, but only 34% say they are genuinely reimagining how the business works, versus 37% who admit AI is still being used superficially, bolted onto existing processes rather than replacing them.
The pattern by company size is stark too: 54% of firms with over $1 billion in revenue report enterprise-scale AI adoption, against 33% of smaller organisations — a gap that tracks budget for integration and change management more than access to the technology itself. In UK financial services, where the return is easiest to measure, the clearest ROI sits in specific, well-scoped use cases: 33% of firms use AI for fraud detection and 37% for cybersecurity, both categories with a direct, quantifiable cost avoided per incident (Bank of England / FCA).
Agentic AI: From Pilot to Production
Agentic AI — systems that plan and act toward a goal rather than simply responding to a prompt — is the clearest dividing line in this year's data. McKinsey found that 40% of large enterprises are now scaling AI agents, up sharply from 27% the year before, compared with just 22% of smaller organisations, a figure that hasn't moved. 20% of organisations have reached the scaling phase specifically with software coding agents, and 32% report they decided to build software in-house using AI tools rather than buy it — a direct signal of growing confidence in owning the build.
Governance has not kept pace with that confidence. Deloitte's 2026 report found that only one in five companies (20%) has a mature model for governing autonomous AI agents, even as usage is "poised to rise sharply" across financial services, aviation, manufacturing and the public sector. That gap is precisely where a badly scoped agentic rollout turns into a governance incident rather than a productivity win.
2025 vs 2026: The UK AI Transformation Scorecard
Metric2025 position2026 positionOrganisations using AI in ≥1 function~78% (McKinsey)89%Enterprise-scale AI adoption38%44%Reporting positive EBIT impact~39%37% (largely flat)Large enterprises scaling AI agents27%40%UK financial services firms using AI58% (2022 baseline)75%Mature agentic AI governance in placeNot separately tracked20% (Deloitte)Challenges Still Holding Transformation Back
Three obstacles show up consistently across this year's sources:
1. Proof-of-concept fatigue
Gartner's oft-cited prediction that at least 30% of generative AI projects would be abandoned after proof of concept by the end of 2025 — citing poor data quality, unclear business value and escalating costs — still describes most failed UK deployments encountered by delivery teams.
2. Governance lagging usage
With only 20% of enterprises reporting mature governance for autonomous agents, most UK organisations are running agentic pilots faster than they can audit them — a gap regulators including the ICO, under UK GDPR, are increasingly likely to test.
3. AI-enhanced cyber risk
The NCSC's assessment of AI's impact on the cyber threat is unambiguous: AI will "almost certainly increase the volume and heighten the impact of cyber-attacks" over the coming two years, primarily by making phishing and social engineering more convincing and lowering the skill bar for less capable attackers.
Building a Transformation Roadmap That Survives Contact with These Numbers
The businesses closing the adoption-to-ROI gap tend to do three things differently.
First, they scope AI projects around a single measurable outcome rather than a general “AI strategy”, which is what keeps a pilot out of Gartner's abandonment statistic.
Second, they integrate governance and audit trails from the ground up, rather than adding them on top of an agent already processing customer information — the standard practice of any Companies House-registered supplier.
Third, they treat this as due diligence, not a single purchase: briefing a specialist AI development company UK teams can hold accountable for results will close the productivity gap more quickly than adding another licensed tool to an already-crowded stack.
Frequently Asked Questions
1. What is AI transformation, and how is it different from AI adoption?
Adoption is about using AI technologies; transformation is about reimagining work as a result of them. The 2026 data shows adoption is nearly universal (89%) while true transformation — redesigned workflows with measurable EBIT impact — sits closer to 34-37%.
2. Why isn't AI adoption translating into profit for most UK businesses?
Most deployments are limited – they are applied to existing processes and don't replace them. Both McKinsey and Deloitte report that a very small proportion of organisations have reimagined fundamental workflows based on AI, where the financial value really comes into play.
3. Is agentic AI ready for production use in UK businesses?
Partially. Large enterprises are scaling agents fast (40%, up from 27%), but governance maturity for autonomous agents sits at only 20%, so production use should come with an audit and oversight framework, not just a pilot budget.
4. What's the single biggest risk in an AI transformation programme right now?
Cyber risk compounded by governance gaps. The NCSC expects AI to increase both the volume and impact of attacks over the next two years, at the same time as most organisations lack mature controls for the AI agents now handling sensitive processes.
Conclusion
The 2026 data settles one argument and opens another. UK businesses no longer need convincing to adopt AI — 88% already have. What separates the organisations seeing real EBIT impact from the 63% still waiting is scope discipline, governance built in from the start, and a genuine build-versus-buy decision rather than a stack of disconnected tools. That's also where the case for a proper AI consulting UK engagement holds up: not to sell another pilot, but to turn this year's adoption numbers into next year's ROI numbers.
Businesses weighing up that next step can find more detail on RSK Business Solutions' AI services.
