The headline everyone reaches for is job losses. New data from two independent sources this week tells a different, more useful story: AI is creating jobs, not destroying them at scale, but the businesses actually benefiting aren't the ones with the most access to the tools. They're the ones that have done the work on skills, culture and confidence.

The Lloyds Business Barometer surveyed 1,200 UK companies in July and found 54% of employers say AI has already led to job creation in their business, with a quarter increasingly hiring specifically for AI skills and a fifth creating AI-specific roles outright. That's a genuinely upbeat number, and it deserves to travel further than it currently does.

Sit inside that headline for a moment and a sharper picture appears. Businesses with turnover above £10 million are far more confident their workforce already has the AI skills it needs, roughly two-thirds of them say so, compared with just over half of firms overall. Same survey, same technology, two very different levels of confidence. The difference isn't size for its own sake. It's what larger firms have typically already invested in: structured skills development, clear internal ownership of AI, and time given to build genuine capability rather than assuming a licence is the same thing as capability.

Amanda Murphy, Lloyds' CEO of business and commercial banking, put the shift precisely: the challenge now is moving "from accessing AI to building the skills, culture and confidence to use it effectively." That sentence is worth reading twice, because it names exactly what separates a business that has bought some software from one that has actually changed how it works.

A second piece of research published this week adds useful context. Apollo's chief economist Torsten Slok analysed hundreds of occupations and found that jobs with higher AI exposure are showing weaker wage growth, while the overall employment effect remains small and, in his word, insignificant so far. AI is also fuelling record levels of new business formation, which points to job creation still working its way through the economy over time rather than arriving as a one-off spike.

Put the two studies together and the honest read for a smaller business is this: the risk right now isn't mass redundancy. It's pay and opportunity quietly drifting toward businesses that took AI adoption seriously early, while everyone else's teams do broadly the same work for broadly the same money, a year or two behind where they could be.

This is exactly the pattern that shows up on the ground. A 36-person events and film agency recently went through a full AI readiness diagnostic before a single training session was designed, precisely so the training landed on the actual gaps rather than a generic curriculum. The result was a six-session programme that scored 4.6 out of 5 across 131 survey responses, with not one dissatisfied response. The tools the agency already had access to didn't change. What changed was the skills, the culture around using them, and the confidence to do it properly.

That's the real dividing line the Lloyds data is describing. It isn't access to AI. Most businesses now have that, in some form, whether they planned for it or not. It's whether anyone has done the deliberate work of building capability around it, so the tool actually gets used well rather than tentatively, inconsistently, or not at all.

For a business trying to work out where it sits, the honest starting point isn't another tool comparison. It's a clear-eyed look at where the business's own skills, culture and confidence stand today, and what the realistic next step looks like. That's what an AI Accelerator Diagnostic is built to answer: a half-day assessment across six dimensions, ending in a written roadmap and one immediate action.

The jobs story is better than the headlines suggest. It just isn't evenly distributed, and the businesses that get ahead of that will be the ones that decided to, not the ones that waited to find out.

Go wisely.