Key points
  • 90% of UK-based respondents said their firm already uses AI in most processes — seven points ahead of United States respondents.
  • The sample was 1,550 enterprise AI decision-makers at more than 500 organisations with revenues over $100 million, across six markets.
  • 53% said AI is fully embedded in core workflows; only 19% of UK respondents called their AI environment “fully integrated”.
  • The top barrier named was integration across legacy systems (38%) — a problem most micro-businesses do not have.
  • Publicis Sapient's own global announcement of the same research led with different figures: 73% using AI regularly, and 10% calling it core to operations.

Research from Publicis Sapient, reported by Consultancy.uk on 28 July 2026, found that 90% of UK-based respondents said their organisation already used artificial intelligence in most processes — seven percentage points ahead of their United States counterparts. Headline figures of this kind are frequently recycled as evidence that smaller firms are falling behind. The methodology is worth reading before drawing that conclusion.

Who was actually surveyed

Consultancy.uk reports that the responses came from 1,550 enterprise AI decision-makers at more than 500 global organisations with revenues over $100 million, based in six markets: the United Kingdom, the United States, France, Germany, the United Arab Emirates and Australia.

That is a survey of large enterprises. It is a legitimate and useful survey of large enterprises. It is not a measurement of AI adoption among UK sole traders, micro-businesses or firms with fewer than ten staff, and it does not claim to be.

Worth knowing The single most-cited barrier to scaling AI was integration across legacy systems, at 38%. Among UK respondents specifically, integration across business units (41%) and governance (36%) led the list. A five-person business has no legacy system estate and no cross-unit silos — so the headline constraint in this research does not transfer downward.

What the researchers said

“Organisations are under pressure to move faster on AI, and while businesses are increasingly embedding AI into workflows, that momentum stalls if large portions of IT budgets remain focused on maintaining legacy systems. Modernising these environments to strengthen digital foundations is critical to supporting the next phase of AI adoption. Without that groundwork in place, scaling AI across the organisation becomes far more difficult.”

— Harshu Deshpande, Vice President, Publicis Sapient, quoted by Consultancy.uk, 28 July 2026

Deshpande's point is about IT budgets and legacy modernisation. It is an enterprise argument, addressed to enterprises.

The figure the same research reported differently

This is worth flagging, because the story is more nuanced than one number. Publicis Sapient's own announcement of the 2026 Global Enterprise AI Report led with a different framing of the same study: 73% of respondents said AI is used regularly or across most business processes, while only 10% said AI is core to how their business operates. Nigel Vaz, Chief Executive of Publicis Sapient, said in that announcement:

“The enterprise was not designed for the speed, scale and autonomy that AI makes possible. Many organizations have successfully deployed AI, but deployment alone does not create advantage.”

— Nigel Vaz, CEO, Publicis Sapient, in the company's 2026 Global Enterprise AI Report announcement

The 90% and the 73% are different cuts — one UK-only, one global, and they measure different questions. But the gap between “uses AI in most processes” and “AI is core to how the business runs” is the substance of the report, and it tends to disappear when the headline travels on its own.

A more useful question for a small UK business

Enterprise adoption rates do not tell a small business anything actionable. A sole trader is not behind a $100 million company; they are solving a different problem with different constraints.

The question that does produce an answer is narrower: does one specific tool save more hours per month than it costs? That is arithmetic — subscription cost against hours recovered, valued at what an hour of your time is worth. It requires no view on whether AI is transformational.

  • Pick the single most repetitive admin task in the week.
  • Time it honestly over a fortnight.
  • Price a tool that addresses that task specifically.
  • Compare the monthly cost against the hours returned. Stop if the maths does not work.

Common questions

Do 90% of UK businesses really use AI?

90% of respondents in a Publicis Sapient survey said their organisation uses AI in most processes. Those respondents were 1,550 AI decision-makers at organisations with revenues above $100 million. The figure describes large enterprises surveyed, not all UK businesses.

Who carried out the research?

Publicis Sapient, a digital consultancy. The UK findings were reported by Consultancy.uk on 28 July 2026 as part of the firm's 2026 Global Enterprise AI Report.

Does this mean a small business is behind on AI?

Not necessarily. The survey covered organisations turning over more than $100 million, and the leading barrier it identified — integration across legacy systems — is not a constraint most small firms face. Enterprise adoption rates are a poor benchmark for a business with fewer than ten people.

What is the difference between the 90% and 73% figures?

They are different cuts of the same research answering different questions. The 90% is UK respondents saying AI is used in most processes. The 73% is the global figure for regular or widespread use, published alongside a finding that only 10% consider AI core to how their business operates.

How should a small business decide whether to adopt an AI tool?

By comparing the cost of a specific tool against the hours it returns on a specific task, rather than against industry adoption statistics.

In short

A Publicis Sapient survey found 90% of UK respondents at organisations turning over more than $100 million use AI in most processes. The sample size is 1,550 enterprise decision-makers and the leading barrier identified is legacy system integration. Neither the population nor the barrier describes a typical UK small business. The useful test remains whether one tool saves more hours than it costs.

Based on reporting by Consultancy.uk, 28 July 2026.