UK businesses overstate AI progress, EXL research finds

EXL has published research suggesting that most UK businesses overstate their progress in adopting artificial intelligence. The study found that only 12% of companies met its definition of an AI Leader.

The survey of 212 C-suite and senior decision-makers in banking and finance, insurance, retail, and utilities pointed to a gap between executive confidence and company-wide AI implementation.

Three-quarters of respondents said their organisations were ahead of competitors on enterprise AI integration. Yet few had moved beyond limited deployment to broad use across core business functions with measurable returns.

According to the research, 38% of UK companies had moved agentic AI beyond the pilot stage. Even so, scaling its use across an organisation remains uncommon.

Budget gap

The report also highlighted a sharp mismatch between ambition and spending. While 93% of UK companies said scaling AI was important, only 3% expected to receive incremental budget for that work, compared with 60% of US companies in the research.

Data was identified as the main obstacle to more effective AI use. Overall, 77% of respondents said data issues were their biggest challenge, even though better data quality and accessibility were also seen as the most important step in expanding AI use.

Companies succeeding with AI were about five times more likely to report having leading-edge data management programmes than those falling behind. The finding points to a practical divide between experimentation and broader operational use.

Sector results varied widely. Utilities reported the strongest financial gains from AI and agentic AI, while insurers recorded the weakest.

Utilities in the UK reported average improvements of 27% from AI and 28% from agentic AI across cost reduction, revenue growth, and margin expansion. Insurance companies reported average returns of 18% and 17%, respectively.

The research linked some of that difference to sector priorities. Among utility organisations, 61% rated scaling AI as extremely important, compared with 33% of insurers.

Operating model

The report argued that the companies making the most progress are reorganising their businesses rather than simply adding AI to existing structures. The most advanced users had embedded AI into higher-impact workflows and changed how work is carried out.

That shift was reflected in responses on operating model redesign. Last year, 52% of companies classed as AI Leaders said they had completely redesigned their enterprise-wide operating model to deploy and benefit from AI. This year, 35% said they had done so.

Among companies classed as AI Laggards, the share reporting an operating model redesign rose to 38% from 3% a year earlier. The figures suggest slower movers are beginning to make broader structural changes as pressure grows to produce clearer returns from AI spending.

Bhupender Singh, President and Head of International Growth Markets at EXL, said the gap between perception and execution is becoming harder to ignore as businesses face pressure to deliver returns.

"As the pressure mounts for companies in the UK to deliver measurable results from AI, leaders in this space are distinguishing themselves by reimagining their business from the ground up," said Bhupender Singh, President and Head of International Growth Markets at EXL.

"In order to catch up, companies that are lagging behind need more than just technology; they need to redesign their operating models, from workflows to talent management and underlying data infrastructure," Singh said.