Every large UK company now has some form of AI strategy, and every earnings call features some version of an AI update. The harder question — what’s actually changing inside these organisations, in ways that matter beyond a slide deck — has a more specific and more mixed answer than either the enthusiastic or alarmist coverage tends to suggest.
The productivity case is real, but narrower than the pitch
Bank of England analysis has found that companies genuinely adopting AI are seeing measurable productivity gains — software and IT consulting contributed roughly 0.1 percentage points to UK productivity growth between 2023 and 2025, around ten times the equivalent contribution during the pre-Covid decade. That’s a real, measurable effect, not just anecdote. But it’s concentrated in specific functions — software development, data analysis, certain categories of customer service and content production — rather than uniformly transforming every part of a large organisation at once, and the same Bank of England analysis noted that these productivity gains are, in a meaningful share of cases, coming alongside reduced headcount in the affected roles rather than purely additive output.
The jobs effect is already visible in hiring data, not just speculation
This is the part of the AI-and-work story that tends to get softened in corporate messaging but is genuinely showing up in labour market data: UK job postings for occupations with high exposure to AI automation fell 38% between 2022 and 2025, compared with a 21% decline for lower-exposure roles — a meaningful gap that points to AI adoption actively reshaping hiring, not just productivity within existing roles. Separate analysis has found that around one-fifth of UK tasks are AI-exposed, a higher share than any of eleven other advanced economies studied in the same research, with London’s concentration of finance, legal and media white-collar work making the capital particularly exposed to this shift.
The counter-evidence: this isn’t a one-way story
It would be inaccurate to present this purely as a jobs-destruction narrative. A 2026 study found 54% of UK businesses reporting that AI adoption has created new jobs within their organisation, even as it’s changed or eliminated others — new roles in AI oversight, prompt engineering, data governance and model evaluation didn’t exist in most companies a few years ago and now do. The honest picture is a labour market being reshaped rather than straightforwardly shrunk: net headcount effects vary considerably by sector and company, and there’s no reliable evidence yet of the kind of mass-scale replacement that the most alarmist commentary sometimes implies.
Where the change is actually happening inside companies
Talking to people inside large UK organisations rather than reading their public AI messaging reveals a more mundane picture than “AI transformation” suggests: the most consistent, genuine change is in specific, bounded tasks — first-draft content generation, code review assistance, customer query triage, contract review — being handled with AI assistance rather than fully manually, with human oversight remaining the norm rather than the exception in most functions that matter for decision-making. Wholesale replacement of entire departments or decision-making functions remains rare outside a small number of aggressive early adopters, and several high-profile corporate AI rollouts have been quietly scaled back after overpromising on what the technology could handle unsupervised.
What boards are actually being told, versus what they’re saying publicly
There’s a genuine gap worth noting between external corporate AI messaging — often ambitious, transformation-focused language aimed at investors — and the more cautious, incremental reality board members are reportedly being given internally, where risk, implementation cost, and workforce management concerns feature far more prominently than the public narrative suggests. That gap isn’t necessarily dishonest; it partly reflects the different audiences and purposes of investor communication versus internal planning, but it’s a reasonable reason for outside observers to treat sweeping corporate AI claims with some scepticism until the underlying financial and operational results confirm them.
How company size changes this picture considerably
The evidence and examples discussed so far skew heavily toward large, well-resourced companies, and it’s worth being direct that the picture looks meaningfully different for smaller UK businesses. Large companies generally have the capital, technical expertise and risk tolerance to run careful, well-governed AI pilots with genuine oversight built in. Smaller UK businesses, adopting the same underlying AI tools, often do so with considerably less structured implementation — sometimes little more than individual employees using consumer-grade AI tools without formal company policy, data governance, or oversight at all. This means the productivity and job effects described above, drawn largely from research and reporting focused on larger organisations, may understate the more chaotic, less measured version of AI adoption happening simultaneously across the much larger population of small and medium UK businesses, where the effects are considerably harder to study systematically but plausibly just as significant in aggregate.
What employees inside these companies are actually experiencing
Beyond the productivity statistics and headcount data, it’s worth briefly noting the more subjective, harder-to-quantify experience of working inside a company actively adopting AI tools. Employee surveys and workplace research point to a genuinely mixed picture: some staff report AI tools removing tedious, low-value parts of their role and freeing time for more engaging work, while others report increased pressure to demonstrate productivity gains that AI adoption is meant to deliver, alongside genuine anxiety about job security that persists even at companies publicly emphasising augmentation over replacement. Both experiences are real and are being reported simultaneously within the same organisations, often by employees in similar roles, which is itself a reminder that “how AI is affecting work” doesn’t have a single, uniform answer even within one company.
What this article is not
This is a description of documented trends in UK corporate AI adoption and its labour market effects, not investment advice or a prediction about any specific company’s future performance. AI’s effects on any given business, sector or job are genuinely uncertain and evolving, and this piece should be read as a snapshot of the current evidence rather than a settled long-term forecast.
Sources: Bank of England analysis on AI adoption and productivity; UK job-posting exposure research; 2026 survey research on AI and UK business job creation.
