BCG’s fourth ‘State of UK Business’ survey from the Centre for Growth comes at a timely point for UK politics and economics. Since our 2025 survey was published in March last year, we’ve seen a new government, renewed concerns over energy prices and inflation from the conflict in the Middle East, plus continued subdued consumer spending.
Against this backdrop, we surveyed over 2,300 UK business leaders to understand the impacts on their businesses, where they see opportunities and their future plans. This year’s survey also included a deep dive on AI, examining how businesses are using it, whether they are capturing value from the technology and what separates the AI leaders from the laggards.
At first glance, business leaders appear to be slightly more optimistic than last year: 87% describe themselves as confident about their business’s prospects over the next 12 months (+4pp from 2025 level) (Exhibit 1.1).
However, looking past the headline numbers, this positivity is not backed by confidence in the fundamentals. Compared with last year, fewer leaders expect their profits to grow (48%, -8pp from 2025) and more expect them to decline (28%, +10pp) (Exhibit 1.2).
When it comes to workforce size, the balance has also tipped. A third (33%) of businesses now plan to shrink headcount (+10pp), narrowly overtaking the 31% planning to grow it, which is unchanged on last year.
Recession fears have not improved from last year (57%, flat) and even the majority (57%) of leaders who are confident about the year ahead still expect a recession. It is also now the biggest businesses (60%, +6pp) and multinationals (61%, +5pp) that are most convinced a recession could happen. This is not inconsequential: these businesses have the greatest visibility into global markets and tend to be least bearish on this front. If this type of firm is more likely to think a recession could happen in the UK over the next year, it is a worrying signal for UK business confidence more broadly.
Taken together, this suggests something other than a genuine turnaround in business confidence. Where optimism is rising, it looks to be driven less by conditions improving but rather because 2025 was a particularly poor year. Many leaders appear to be reporting confidence in their ability to cope with challenges, rather than confidence that the year ahead will be better. That is a thinner, more precarious, form of optimism than the headline number implies, and is not an especially encouraging signal for an economy still struggling with growth.
That distinction between what leaders say at the top line, as opposed to what they actually plan to do, sits at the heart of this year's findings and is easiest to see once businesses are grouped by their own growth expectations rather than treated as one block. Businesses expecting to grow are doing the most: they see more opportunities, and they are acting on more of them, especially where technology is involved. Across the whole sample, the three opportunities mentioned most often were investing in new technology, using everyday AI tools and deploying agentic AI. Growth-expecting firms are far more likely to be pursuing all three than businesses expecting to shrink, which instead tend to look outward, hoping for tax changes, lower costs, or lighter regulation to improve their position (Exhibit 2.1).
There is a third group, however, that stands out more than either of these and arguably tells the bigger story. Businesses that expect to stay roughly the same size, whether in terms of profit or by headcount, are the least engaged group of the three across the board: they are the group least likely to identify an opportunity, a threat or a barrier at all (Exhibit 2.2).
It is a phenomenon that aligns with our separate research on UK productivity. This showed the tail of low-productivity businesses in the UK has lengthened and widened over the past two decades, even as its top performers held firm. Put the two together and a similar picture emerges here: an active group of growth-oriented firms pulling ahead, decline-expecting firms chasing efficiency measures, but a much larger group sitting still.
That stagnant middle, far larger in number than the frontier firms, is arguably where UK growth policy needs to focus if it is to address stagnant productivity growth.
When it comes to what leaders want from government, tax has topped the policy wish list every year we have run this survey. 2026 is no exception, although the strength of feeling has cooled a little. Business taxes (33%, -9pp), energy costs (27%, -10pp) and employment taxes (25%, -3pp) are still the three most-cited priorities, and tax reform is still the single change leaders think would do the most for the economy (Exhibit 3).
There is some improvement in how leaders see the government's direction. Half (50%, +6pp) now think it is prioritising the economy, a meaningful jump on last year. That has not translated into trust, however: 64% (-3pp) still rate the government's overall impact as negative, 66% think its spending priorities are wrong and 73% say policy uncertainty is denting business confidence. Leaders' confidence has also moved a little in the past three months alone (52% feel more confident now than they did then), which may reflect the fact that a new government is still being sized up rather than a definitive view.
If there is one area where the gap between what leaders say and what the data supports is starkest, it is AI.
Sentiment here looks unambiguously positive. Almost three in four leaders (72%) are positive about AI personally, 34% of which are strongly positive. The share that sees it benefiting their own business has risen every year we have asked since 2024 and now stands at 75% (+16pp from 2025). But looking closer, the same divide that runs through the rest of the survey reappears: this optimism is heavily concentrated among people who are already furthest along with the technology. C-level leaders are considerably more positive than founders (84% vs. 58%), and daily users are far more positive than occasional ones (95% vs. 19%). Greater AI usage and engagement leads to more positivity about the technology, not less. Growth-expecting firms also outscore decline-expecting firms on every AI measure we tested (Exhibit 4).
Growth-expecting firms are considerably more likely than their peers to point to quality and productivity gains as reasons for using AI, while the two groups converge closely on every other reason. Over half (54%) of growth-expecting firms cite improved quality of products or services, against just over a third (36%) of firms expecting flat profit and 31% of firms expecting profit to decline, a gap of 18-23pp (Exhibit 5.1). Across the remaining reasons tested, the three groups sit far closer together. This suggests growth-expecting firms are not simply more active users of AI; they are using it with a clearer sense of the business outcomes they expect from it.
We also tested how leaders would react to a hypothetical windfall: a 30% saving in time from using AI. Growth-expecting firms said they would put it toward expanding into new markets or reinvesting in more AI. Decline-expecting firms were twice as likely to say they would use the same gain to cut headcount or freeze hiring instead (Exhibit 5.2).
Self-perception adds another layer to this. More than half of leaders (53%) believe their business is ahead of competitors on AI, while only one in ten (10%) admit to being behind, a split that simply cannot be true as a whole. It suggests a good number of leaders are misjudging where they actually stand. This is a problem if policy support depends on businesses first recognising they need help: it risks flowing to those already ahead while passing over the businesses that would gain most from it.
The obstacles each group reports are consistent with this. Businesses that consider themselves as ahead on AI point to practical, solvable problems: security, cost, regulation and finding the right skills. Businesses that consider themselves behind point to something more basic: they are not sure what AI would actually do for them, where to begin, or even whether their current way of working needs to change at all. Cost, despite the attention it gets externally, is cited by fewer than a third of businesses overall, which suggests that by the time cost becomes a point of interest, most firms have already concluded AI is worth having.
That gap looks even wider among businesses not using AI at all. More than half (53%) say the idea has simply not crossed their mind. When asked why, the two most common answers were that AI is not relevant to their line of work (47%) and a preference for keeping things human (45%); cost trailed well behind at 12% (Exhibit 6). On the relevance point specifically, our own analysis of the wider economic opportunity from AI put the baseline gain from current basic AI tools at 14% across broadly transferable tasks across the economy. This suggests a ‘not relevant to us’ view may not stand up as well as leaders think.
Put all this together and a single thread runs through this year's survey: an economy that is divided. Business confidence has genuinely risen, but it is confidence built on having made it through another challenging year rather than on any real conviction that growth is coming. Underneath it, a smaller group of ambitious, tech-forward businesses are separating from a much larger group that is neither growing nor shrinking, but just holding on. AI is seen as a genuine opportunity not only to drive efficiencies but achieve growth, although the gap between firms capturing value and growing and the rest is already stark and likely to widen further without targeted intervention. The practical challenge for government and industry alike is to make sure the vast majority of firms don’t get left behind.
About Us
BCG's Centre for Growth brings together ideas, people and action to drive the UK forward. We work with our global expert network to identify transformational opportunities, connect key decision-makers and build coalitions for change. We offer long-term strategic insight, extensive cross-sector expertise, platforms for dialogue and bias to action.