Solving Europe's AI adoption puzzle

A new report from London Business School and European tech, startups and venture capital news portal, Sifted ,finds that Europe's AI challenge is less about technology than organisation - and that the region's path to competing may lie in trust, not scale.

Europe has the research base, the capital and the policy ambition to lead in AI. So why is adoption lagging?

That's the question at the heart of a new Sifted report, produced in partnership with London Business School, which argues that Europe's AI story is not one of technological deficit but of organisational deficit.

The adoption gap

Around 43 per cent of US workers use AI in their jobs, against a European average of 32 per cent. Just 20 per cent of EU companies use AI for any purpose, compared with 34 per cent in the US. Even globally, McKinsey data shows that while 88 per cent of organisations use AI somewhere in the business, only 7 per cent have fully scaled it.

"There is no inherent roadblock to Europe leading in AI adoption," said Nicos Savva, Professor of Management Science and Operations and Academic Director of the School's Data Science and AI Initiative. "We have the technical capital and the ability." The bottleneck, he argues, is organisational: management, workflow redesign, skills, incentives and trust.

Why diffusion is slow

Executives interviewed for the report point to a "people element" that was widely underestimated. Concerns range from unreliable AI outputs and cybersecurity risk to genuine anxiety about job losses. Lynda Gratton, Professor of Management Practice, warned that leaders who frame AI as a threat to jobs undermine their own adoption efforts: "If you feel AI will destroy your work, then you won't want to adopt it."

Others flagged a subtler risk: Isabel Fernandez-Mateo and London Business School's Dean Sergei Guriev both cautioned against "cognitive atrophy" - a wholesale delegation of thinking to AI that could erode the skills organisations will need long-term.

Scaling Europe's champions

The report also examines why European AI companies struggle to scale, pointing to the region's compute disadvantage and shallower capital markets. But Gary Dushnitsky, Deputy Dean and Professor of Strategy and Entrepreneurship, suggests Europe's opportunity lies elsewhere: building trusted, embedded, domain-specific AI rather than competing head-on for frontier models. "Perhaps the European value proposition should be built around resilience, confidentiality and privacy," he said.

The policy lever

On regulation, the report finds a divided picture. Some argue that sandboxes, smart procurement and skills investment can accelerate diffusion; others, including Ekaterina Abramova, call for tighter global oversight, particularly around military and agentic AI applications. Keyvan Vakili predicts a "big re-sorting" of the labour market over the next decade, even if a dramatic "jobpocalypse" fails to materialise.

The bottom line

The report's conclusion is deliberately balanced: Europe doesn't simply need more AI - it needs better AI adoption, embedding the technology in ways that boost productivity while protecting the human capabilities that make it work.

Read the full report, Solving Europe's AI Adoption Puzzle, produced by Sifted in partnership with London Business School's Data Science and AI Initiative.