Key takeaways
- 54% of firms say their workforce has the AI skills they need. 31% say it does not. Among firms above £10m turnover it is nearly two-thirds.
- 61% of firms use AI. Among firms above £10m it is 79%. Adoption itself splits by size before skills even come into it.
- 77% of larger firms plan to increase AI investment, against 58% of firms overall. The leaders are accelerating.
- 54% said AI created new jobs at their firm. That is the number to put in front of anyone still assuming AI only cuts headcount.
- The commercial point: the grant rules are more generous to non-levy employers, not less. The barrier for smaller firms is attention, not eligibility.
The usual way to read an AI skills survey is by sector: tech is ahead, manufacturing is behind, the public sector is somewhere in the middle. The Lloyds Business Barometer, published on 18 August 2026, suggests a more uncomfortable dividing line, and it runs straight through the middle of the British economy.
The finding
| All firms | Firms above £10m turnover | |
|---|---|---|
| Currently use AI | 61% | 79% |
| Workforce has the AI skills needed | 54% | Nearly two-thirds |
| Plan to increase AI investment | 58% | 77% |
| See failure to adopt AI as a competitive disadvantage | Nearly 6 in 10 54% of firms under £1m | 73% |
Source: Lloyds Business Barometer, published 18 August 2026. Fieldwork of 1,200 UK companies conducted in July, per Bloomberg's reporting of the same survey.
Read down the right-hand column. Larger firms are ahead on adoption, ahead on skills, investing harder, and more convinced it matters. Every row points the same way, and the last one is the most telling: 73% of firms above £10m think failing to adopt AI is a competitive disadvantage, against 54% of firms under £1m. The businesses most at risk from falling behind are the least likely to believe they are.
The organisations that already have the skills are the ones accelerating. A gap where the leaders are also the ones investing hardest does not narrow on its own.
Why size, and not sector
The explanation is duller than the headline and more useful.
Large firms have machinery for this. A learning and development function. A training budget that survives a bad quarter. Somebody whose actual job is workforce capability, with the authority to plan two years out.
Smaller firms mostly have enthusiasm. AI capability arrives through whoever happens to be curious, usually on a personal account, usually undocumented, usually without anyone senior knowing how far it has gone. That produces real skill, and it produces it unevenly, invisibly and with no way to prove it or build on it.
It is worth being clear that this is not a criticism of smaller employers. It is a structural difference. A 60-person business does not have a spare head to run capability planning, and pretending otherwise is how well-meaning advice gets ignored.
It also explains why hiring is the wrong first instinct if you are small. A quarter of firms are increasingly recruiting for AI skills, and PwC puts the UK wage premium for those skills at 34.2%. Paying that means bidding against exactly the larger, better-capitalised firms in the right-hand column above. That is not an auction a 60-person business wins. We set the arithmetic out properly in the build-or-buy piece.
The other numbers, including one worth quoting at your board
54%
of firms said AI has led to job creation at their organisation
43%
are introducing AI skills training, and 32% expanding what they already run
42%
of those increasing spend expect to commit £25,000 to £100,000
That first figure is the one to keep. The default boardroom assumption is still that AI adoption means headcount reduction, and it is quietly stopping a lot of sensible investment. More than half of surveyed firms report the opposite happening in their own business.
The honest counterweight
We are not going to present that as the whole picture, because it is not.
Hiring for AI-exposed roles such as software developers and consultants has slumped since 2022. The view that AI will ultimately mean fewer jobs is now mainstream among chief executives. And analysis by Bank of England staff has found that firms using AI are becoming more productive, but that those productivity gains are coming partly at the expense of jobs.
Both things are true at once. Firms that adopt AI well tend to create roles. The wider labour market is still shedding them. If you are running a business, the first fact is the one you can act on. If you are advising a school leaver, the second one matters more.
What Lloyds actually said the problem is
Firms need to shift their focus from accessing AI to "building the skills, culture and confidence to use it effectively".
Amanda Murphy, CEO of Lloyds Business and Commercial Banking — Lloyds Business Barometer, 18 August 2026Read that as a diagnosis rather than a platitude. Access is a procurement problem, and most firms solved it eighteen months ago by buying licences. Skills, culture and confidence is a development problem, and almost nobody has solved that, because it does not have a purchase order attached to it.
It also explains a pattern we see constantly: an employer with a full Copilot rollout, a low usage rate, and a leadership team quietly wondering whether AI was overhyped. The tool was never the missing part.
Lloyds are not just reporting this. In the same release they set out their own answer: more than 1,000 AI-related roles in 2026, and one of the first Level 6 AI Engineering apprenticeships run by a UK bank. The organisation publishing the research is closing its own gap with apprenticeships.
The bit that should annoy you, if you run a smaller firm
Here is the thing that makes the size gap avoidable rather than inevitable.
The grant rules run the other way. To be precise about it: levy payers have a dedicated pot of their own money to spend, which is a real advantage. But the grant position for a non-levy employer is more generous, not less, and that is the part smaller firms tend not to know.
| Non-levy employer | Levy payer | |
|---|---|---|
| Apprentice aged 16 to 24 | 100% government funded | Drawn from your levy account |
| Apprentice aged 25 or over | You co-invest 25% | From your account, then 25% once it empties |
| Extra incentives | £2,000 hiring payment for eligible young apprentices from October | None equivalent |
| Minimum company size | There isn't one. Existing employees can be apprentices. | |
So the smaller employer with the bigger skills gap is not priced out of closing it. The barrier is attention, not eligibility. What the large firm really has is someone whose job it is to notice. There is a dedicated SME AI apprenticeships page if that is where you sit.
Where a smaller employer actually starts
Not with a strategy document. With one process and one person, in that order.
-
Find out what is already being used
Do: ask your team, without blame, which AI tools they use and what for. Expect to be surprised.
Why: in smaller firms informal adoption is almost always further ahead than leadership thinks, and you cannot build on what you cannot see. -
Pick one process, not a programme
Do: choose a single repetitive task that costs real hours every month and make that the target.
Why: general AI awareness training produces enthusiasm and no evidence. One fixed process produces a number you can show the board. -
Check what you are entitled to before you budget
Do: confirm your funding position, including 100% funding for 16 to 24 year olds if you are non-levy, and the £2,000 hiring payment from October.
Why: most smaller firms assume apprenticeships are for large employers and never look. That assumption is the actual barrier.
For a typical 60-person firm the sequence is: audit informal use this week, pick one process next month, then choose the route. Doing it the other way round, picking a programme first, is how training budgets get spent on enthusiasm.
Then pick the route
A single Level 5 AI unit
Two to four weeks, no end-point assessment, £750. Start with AI Strategy & Opportunity if you need to work out where AI belongs, or AI Adoption & Governance if people are already using it and nobody has set any rules. All three units are listed on the AI Leadership Units page.
AI & Automation Practitioner, Level 4
The ST1512 standard, 15 months, up to £18,000 and levy-fundable, with a working automation live on real company work inside the first few months. No coding background required. Available as a Claude, Copilot or Gemini edition.
AI for Senior Leaders
The same Level 4 standard with a free CMI Level 7 Certificate, for an owner or director who wants the capability and a credential that travels.
We will tell you exactly what your business is entitled to, and what one funded route would cost you. On the call:
- Your funding position, levy or non-levy, and what is fully funded versus co-invested
- One process worth targeting first, based on what your team already does manually
- Which route fits, from a £750 unit to a full Level 4
25 minutes, no obligation. If you are already in decent shape we will say so.
Book a conversation →What we would not claim from this survey
- It is self-reported. "Our workforce has the AI skills we need" measures confidence as much as capability, and larger firms may simply be better at believing their own answer.
- It is 1,200 firms surveyed in July. A useful sample, not a census, and the size split is reported as averages rather than a full breakdown.
- Turnover above £10 million is not "large". Plenty of firms above that line are still mid-market businesses with no L&D function at all. The dividing line is blunter than the framing suggests.
The honest summary
The interesting thing in this survey is not that a skills gap exists. Everyone knows that. It is that the gap tracks company size, and that the firms on the right side of it are the ones accelerating hardest.
The part that ought to change behaviour is that the funding system does not follow that pattern. It is more generous to the smaller employer, not less. The difference between the two groups is mostly that one of them has somebody paid to notice.
If nobody in your business has that job, this is the year to give it to someone.
Frequently asked questions.
What did the Lloyds Business Barometer find about AI skills?
Lloyds surveyed 1,200 UK companies in July 2026. Two-thirds of firms with turnover above £10 million said their current workforce had the AI skills needed to meet business demands, compared with just over half of firms overall. More than 70% of large businesses also said they plan to step up AI investment. A quarter of employers are increasingly hiring candidates with AI skills, a fifth are creating new AI-specific roles, and 54% said the technology led to job creation.
Is AI creating or destroying jobs in the UK?
Both, and the answer depends on where you look. In the Lloyds survey 54% of firms said AI had led to job creation at their organisation. At the same time, analysis by Bank of England staff has found that firms using AI are becoming more productive but that those gains are coming partly at the expense of jobs, and hiring for AI-exposed roles such as software developers has slumped since 2022. Job creation inside adopting firms and job displacement across the wider market are not contradictory.
Why do smaller firms have a bigger AI skills gap?
Larger firms tend to have dedicated learning and development functions, formal training budgets, and someone whose job it is to plan capability. Smaller firms usually have none of those, so AI skills are acquired informally by whoever is curious. The Lloyds data suggests that produces a real gap: two-thirds of firms above £10 million turnover say they have the skills they need, against just over half of firms overall, and larger firms are also investing more heavily.
How can an SME fund AI training?
Through the Growth and Skills Levy. Employers that do not pay the levy receive full government funding for apprentices aged 16 to 24, and co-invest 25% for those aged 25 and over. From October there is also a £2,000 hiring payment for eligible young apprentices. Short Level 5 AI units cost £750 each. There is no minimum company size for apprenticeship funding, and existing employees can be apprentices.
Do you need to hire AI specialists or train existing staff?
For most smaller employers, training is the more realistic route. Lloyds found a quarter of firms are increasingly hiring for AI skills, but PwC's 2026 AI Jobs Barometer puts the UK wage premium for those skills at 34.2%, so hiring means competing on price with far larger firms. Training existing people avoids that competition and starts from staff who already understand your business.
What did Lloyds say employers should focus on?
Amanda Murphy, CEO of business and commercial banking at Lloyds, said firms need to shift their focus from accessing AI to building the skills, culture and confidence to use it effectively. That is a capability and behaviour problem rather than a procurement one, which is why buying more licences rarely moves the numbers on its own.
Where should a smaller employer start with AI skills?
Start by finding out what your people already use, since informal adoption is usually further ahead than leadership assumes. Then pick one process worth fixing rather than a general awareness programme. A short Level 5 AI unit suits a leader who needs to make decisions, while the AI & Automation Practitioner Level 4 apprenticeship suits someone who will build and run the automation. Neither requires a coding background.
Sources: all Lloyds figures from the Lloyds Banking Group press release More than half of UK businesses say AI has created new jobs, published 18 August 2026, including the Amanda Murphy quotation. The sample of 1,200 companies surveyed in July, and the Bank of England staff analysis on productivity and jobs, are from Bloomberg's reporting of the same survey via Staffing Industry Analysts. Responses are self-reported. Wage premium from PwC's 2026 AI Jobs Barometer, 15 June 2026. Funding positions reflect the 2026 to 2027 apprenticeship funding rules; confirm your own position before acting. TESS Group provides levy-funded AI apprenticeships and therefore has a commercial interest in the training route.