EXL Surveyed 212 UK Business Leaders. 75% Say They're Ahead on AI. Only 12% Actually Are.

October 23, 2026 · Josh Agusti, Founder, Aiven Consulting Group

Ask a room full of executives whether their company is ahead of competitors on AI, and most will say yes. A new study from EXL puts a number on how often that confidence is justified, and the answer is not often.

EXL's 2026 Enterprise AI Study surveyed 212 C-suite and senior decision-makers across banking, insurance, retail, and utilities companies in the UK. Seventy-five percent of those leaders said their organization is ahead of competitors on enterprise AI integration. By the study's own benchmark, only 12 percent actually qualify as an AI Leader. The other 88 percent sit somewhere between piloting and genuinely stuck, regardless of how confident they sound in a boardroom.

The leaders who stopped leading

The more interesting finding is not the overestimation itself, it is what happened to the leaders who were actually ahead. Last year, 52 percent of EXL's AI Leaders said they had completely redesigned their operating model, workflows, talent structures, and data infrastructure, to run on AI. This year that figure dropped to 35 percent. At the same time, the Laggard group caught up fast on paper: only 3 percent of Laggards reported a full redesign last year, and 38 percent report one this year. The gap between the leading and lagging groups, measured by stated redesign activity, nearly closed in twelve months.

That convergence should not be read as laggards catching up to leaders. It more likely means something else happened. The leaders who moved fastest in the first wave hit the limits of a surface-level redesign, while the rest of the market learned the vocabulary of operating-model redesign without doing the harder underlying work. Saying a process has been redesigned for AI and having that redesign survive contact with real data, real compliance review, and real customer volume are different claims, and this year's numbers suggest a lot of organizations are making the first claim without having tested the second.

Why the gap keeps reappearing in the same place

Ask EXL's respondents what is actually standing in the way, and the answer is consistent. Seventy-seven percent named data as their biggest challenge in using AI effectively, ahead of talent, technology, and budget. That tracks with a related pattern: companies with a formal AI strategy in place are measurably more likely to show results, which suggests the fix starts further upstream than most leadership teams assume. A leadership team can genuinely believe its organization has redesigned how it operates, because the pilots work and the demos are convincing, while the underlying data foundation that would let that redesign hold up at scale never actually got rebuilt.

Where the return still shows up

The study is not uniformly discouraging. Utilities companies in the sample reported average improvements of 27 percent from AI and 28 percent from agentic AI specifically, across cost reduction, revenue growth, and margin expansion, with insurance close behind at 18 and 17 percent. Those are sectors with relatively structured, well-governed data to begin with, which is consistent with the idea that the return on AI tracks the quality of the data foundation underneath it more closely than it tracks ambition or spend.

What this means before the next board update

For a leadership team, the useful move is not to ask whether the company has an AI strategy, almost every company now claims one. It is to ask for evidence that the operating model has actually changed, not just that AI tools have been added on top of the old one. An honest AI readiness assessment will usually surface the same data and process gaps EXL's respondents named, before a board hears about them from a stalled rollout.

The organizations most likely to close this gap are the ones willing to treat operating model redesign as the actual project, with AI as the capability inside it, rather than treating AI adoption as the project and hoping the operating model catches up on its own.

Key Takeaways

EXL's 2026 UK Enterprise AI Study surveyed 212 C-suite and senior leaders and found 75% believe their organization is ahead of competitors on AI, but only 12% actually qualify as an AI Leader by the study's own criteria.
The share of AI Leaders who had completely redesigned their operating model for AI fell from 52% to 35% year over year, even as Laggards making that same claim jumped from 3% to 38%, meaning last year's leaders and today's laggards are converging toward the middle.
77% of respondents named data, not technology or talent, as the biggest obstacle to real AI progress, suggesting most AI strategy gaps are actually data infrastructure gaps in disguise.