Roughly nine in ten companies are yet to see a penny of financial benefit from artificial intelligence, despite a threefold rise in workplace usage over the past two years, according to the head of the world’s largest consulting firm in Britain and Ireland.

Matt Prebble, chief executive of Accenture UK and Ireland, said the disconnect between enthusiastic adoption and measurable returns now ranks as one of the most pressing strategic questions facing boardrooms on both sides of the Atlantic.

“Over the last two years, we’ve seen three times as many people using AI within the workplace, but that individual productivity … that’s not actually yet translating to real company performance,” he said. His verdict echoes fresh Accenture research showing that only one in ten UK organisations has successfully scaled the technology into core operations.

According to Prebble, the failure to extract value has its roots in companies treating AI as a bolt-on rather than reshaping the way they work “across people, process and technology”.

“We found that one in ten companies are really starting to get the productivity flow through to the bottom line, but on the other hand, 90 per cent of companies aren’t,” he said. He remained confident, however, that AI would yet have a “material impact” on businesses prepared to display the “confidence and the willingness to reinvent” how they operate, with the technology at the centre of the redesign.

His warning lands at a moment when chief executives and chief financial officers are sharpening their pencils over AI budgets. Businesses are increasingly questioning whether the sums they are pouring into AI tokens, the basic units used by large language models to read, remember and generate content, are delivering a defensible return. The growing scepticism mirrors a wider pattern of stalling adoption at large enterprises as doubts mount over AI returns.

Andrew Macdonald, chief operating officer at Uber, conceded last week that the ride-hailing and delivery group had yet to observe any direct productivity uplift tied to its rising AI token consumption. “That link is not there yet, right?” he said. By March, Uber had burned through its annual budget for “agentic”, or autonomous, AI, with the link between greater token spend and useful consumer features still unconvincing.

Microsoft has reportedly told some of its staff to switch to its own in-house model rather than third-party alternatives, in an effort to rein in costs. According to Axios, one unnamed company spent $500 million in a single month on Anthropic’s Claude platform after leaving employee usage uncapped.

The mounting cost pressure has emboldened critics of the sector’s hyper-investment cycle. A widely cited MIT study reported by Fortune found that 95 per cent of corporate generative AI pilots were failing to produce measurable returns, prompting renewed warnings of a possible correction in the valuations and business models of the industry’s leading players.

Cultural headwinds are building too. Pope Leo has criticised the AI industry and called for tighter regulation, while graduates at several US college campuses have booed speakers championing the technology. Prebble acknowledged that AI was suffering from “a bit of a brand issue” in the West, “very different to Asia”, with anxiety over job losses and the pace of change clouding the picture.

“You have seen leaders in the market talking around the job dislocation and giving headlines around the impact on early graduate or next graduate jobs, which I think has created some of the fear out there,” he said.

He insisted, however, that equating greater AI adoption with fewer overall jobs reflected a “narrow view” of productivity. “The further we go in this cycle … things will be done differently. And therefore there’ll be different skills and different capabilities required,” he added. “There’s always been those waves of technological change that have come and it is true that it’s always created new job opportunities and over time, those job opportunities have outpaced the previous job.”

For all the gloom over returns, Prebble argued that Britain still has time to turn AI into a national growth story. The UK may have largely missed out on the spoils of building AI infrastructure, but he believes there is a credible path to capitalise on the application layer by playing to British strengths in life sciences and professional services. That view aligns with separate HSBC research suggesting AI adoption could unlock a £105bn revenue boost for UK mid-sized firms by 2030.

“If we can get our innovation swagger back to be able to then scale that across the country and globally, we’ve got some good opportunities,” Prebble said.

Accenture has begun rebranding its 800,000-strong workforce as “reinventors”, a label Prebble said reflects the group’s growing remit advising clients on how to overhaul their operating models for the AI era. Last year the consulting giant restructured its own business, folding strategy, consulting, creative, technology and operations into a single division dubbed “reinvention services”. Earlier this year, reports emerged that the Dublin-based firm had been monitoring how its own staff used AI tools as a factor in promotion decisions.

For now, though, the message from the boss of Britain’s largest professional services consulting brand is blunt: the productivity revolution promised by AI is still, for the vast majority of UK plc, a promise rather than a payslip.

Read more:
Nine in ten companies still waiting for AI to pay off, warns Accenture chief