Good news doesn't get shared as fast as bad news, so it's worth pausing on this one. Lloyds Bank's latest Business Barometer, based on a survey of 1,200 UK businesses, found that 54% say AI has already created new jobs at their organisation, and 61% now use AI in some form. That's a genuinely useful counter to the steady stream of "AI is coming for your job" headlines: for most UK businesses, the story so far is creation, not just displacement.

But the barometer draws a second line that matters more for most readers of this post than the headline number. Adoption splits sharply by size. Among businesses turning over more than £10 million, 79% use AI. Among domestic-only, smaller firms, that figure is 46%. Larger firms are also far more confident in their teams' AI skills, with nearly two-thirds saying their workforce has what it needs, and they're far more likely to see non-adoption as a competitive threat, 73% versus 54% for firms under £1 million turnover.

Read together, those two numbers tell a specific story. It isn't that smaller businesses don't want to use AI, or don't see the point. The barometer names the actual barriers: cost, data quality and access to skills, in that order. Smaller firms aren't held back by scepticism. They're held back by not knowing where to start and not having the internal capacity to work it out alone.

That's a closable gap, not a permanent one, and it's worth being precise about what closes it. It isn't a bigger AI budget or a rush to buy tools. Cost was named as a barrier, but so was data quality and skills access, both of which point at foundations rather than spend. A business that doesn't know what its data actually supports, or hasn't worked out which of its people are ready to lead adoption, will struggle to get value from AI regardless of budget.

What actually helps a smaller business close the gap

The businesses in the £10m+ bracket didn't get there by accident. Somewhere along the way, most of them did the unglamorous groundwork: worked out what AI could realistically do for their specific operations, checked their data and tools were fit for purpose, and built at least a basic plan for governance and capability. That's not a large undertaking. It's a half day spent mapping where a business actually stands, not where it assumes it stands, across the areas that matter: vision, opportunities, people, tools, governance and value.

That mapping is exactly what a structured diagnostic is for. It replaces guesswork with a plan, and the plan is what turns "we should probably use AI more" into "here are the three things worth doing first, in order." For a domestic-only firm sitting in the 46%, that's a more useful next step than either waiting for confidence to arrive on its own or rushing to buy a tool because a competitor mentioned one.

The Lloyds data is a rare thing in AI coverage: a genuinely reassuring number paired with a genuinely useful one. AI is creating jobs, not just cutting them, at least so far. And the adoption gap between larger and smaller firms is about resources and know-how, not appetite, which means it responds to structure rather than persuasion.

If your business is in the 46% rather than the 79%, that isn't a verdict on where you'll end up. It's a starting point, and the fastest way through it is finding out precisely what's missing before spending anything on fixing it.

Go wisely.

Source: Lloyds Business Barometer: AI and jobs, published 18 August 2026, based on a survey of 1,200 UK businesses.