In short: Nearly half of UK businesses that use AI (46%) have given staff no training in it, according to the latest ONS data. Adoption has raced ahead while capability has been left behind, and the gap is widest among the smallest firms. Buying an AI tool is not the same as building the ability to use it, which is why so many businesses see little return. The fix is putting AI inside the systems your team already knows.
There’s no shortage of headlines about British businesses embracing artificial intelligence. The reality on the ground is quieter though, and a lot more awkward. According to the Office for National Statistics, 46% of UK businesses that use AI haven’t given staff any AI training whatsoever.
Not a workshop, not a lunch-and-learn, not a single hour. The tools have arrived. Training hasn’t.
That figure comes from the ONS Business Insights and Conditions Survey, the same national dataset behind our study of AI adoption statistics. It matters because it exposes the weak point in the AI story. Adoption has almost tripled since late 2023, but the skills to use it haven’t kept pace. A business can own the cleverest software on the market and still get nothing back from it if nobody knows how to use it.
The Training Gap in Numbers
The ONS figures show that only about 11% of AI-using businesses with 10 or more employees have trained more than half their workforce on AI. Put simply, meaningful, widespread training is rare. The 46% headline is stark, but the detail underneath is where it stings…
The picture from staff is similar. Recruitment firm Hays found that 58% of UK employees have had no formal AI training or guidance at work. Meanwhile, the ONS finds 55% of employees use AI for work anyway. Most of the people using AI at work, then, are teaching themselves.
How Much of the Workforce Has Been Trained in AI
| Business size | More than 75% | 50% to 75% | 25% to 49% | 1% to 24% | No staff trained | Not sure |
|---|
Percentage of businesses by share of workforce that has received training in the use of AI. Base: businesses that report using at least one AI technology, or are not sure whether they do. Rows may not sum to 100 because of rounding. Not published means the ONS suppressed the cell because the sample was too small. Source: Office for National Statistics, Business Insights and Conditions Survey, Wave 159 (15 to 28 June 2026). Contains public sector information licensed under the Open Government Licence v3.0.
The Smaller the Business, the Worse the Picture
Among AI-using firms with fewer than ten staff, 49% have trained nobody. Step up to the largest firms with 250+ employees and that falls to just 13%.
In other words, the businesses with the least slack, budget, and specialists are also the ones least likely to build any AI capability.
Manufacturing comes off worse still. Over half of AI-using manufacturers (50%) haven’t given any AI training to their staff. That’s a lot of untapped potential sitting idle in an industry under constant pressure to do more with less. It’s also an industry that keeps much of the country running.
This isn’t an isolated finding either. Separate research from across 2026 tells the same story from different angles. Recruitment firm Hays reported that 47% of businesses face moderate or extreme AI skills shortages. Research by Gardiner and Theobald, cited by the British Chambers of Commerce, found 97% of UK organisations report at least one significant AI skills gap.
Why Untrained AI Is Wasted AI
It’s tempting to treat training as a nice-to-have, something to get to once the tools are bedded in. The data suggests the opposite. When AI is dropped into a business without anyone being shown how to use it, one of two things tends to happen. Spoiler: neither is good.
- The first is that the tool simply goes unused. It sits in the corner as an expensive subscription nobody opens, and quietly renews each month. It’s a pattern we see often, where teams aren’t using Copilot properly simply because nobody showed them how.
- The second outcome is more troubling. Keen to keep up, staff start using AI tools on their own, without guidance or guardrails. This so-called shadow AI is now widespread: the ONS finds 55% of employees use AI for work, against 35% of businesses reporting adoption. It brings real risks around data security, accuracy, and compliance.
Either way, the business carries the cost and misses the benefit. The lack of expertise is what holds most firms back, and it’s the most cited barrier in every size band, ahead of cost. Businesses that name a shortage of AI expertise as a barrier are far more likely to be training their people. Some 62% do, against 26% of businesses that report no barriers at all.
Businesses That Feel the Skills Gap Are the Ones Training for It
| How the business is building AI skills | Cite lack of expertise as a barrier | Report no barriers |
|---|---|---|
| Training or retraining existing staff | 62.0% | 26.4% |
| Automating or replacing roles with AI | 12.2% | 17.0% |
| Recruiting new staff with AI skills | 2.6% | 2.2% |
| Outsourcing roles | 6.5% | 4.4% |
Percentage of businesses with 10 or more employees reporting each approach, by whether they cite a lack of AI expertise as a barrier or report no barriers to adoption. Businesses can report more than one approach. Source: Office for National Statistics, Artificial intelligence in UK businesses: 2023 to 2026, Figure 11 (Business Insights and Conditions Survey, Wave 159). Contains public sector information licensed under the Open Government Licence v3.0.
How Businesses Are Building AI Skills, by Size
| Business size | Training or retraining existing staff | Automating or replacing roles | Recruiting new staff with AI skills | Outsourcing roles |
|---|---|---|---|---|
| 0 to 9 employees | 31.7% | 12.0% | 2.1% | 4.7% |
| 10 to 49 employees | 35.6% | 8.2% | 5.4% | 2.5% |
| 50 to 99 employees | 42.2% | 10.2% | 8.0% | 1.5% |
| 100 to 249 employees | 43.0% | 7.6% | 7.0% | 1.5% |
| 250 or more employees | 41.7% | 9.7% | 10.3% | 1.6% |
Percentage of businesses reporting each approach to integrating AI-related skills into their workforce. Businesses can report more than one approach. “Other”, “not sure” and “not applicable” are excluded. Source: Office for National Statistics, Artificial intelligence in UK businesses: 2023 to 2026, Figure 12 (Business Insights and Conditions Survey, Wave 159). Contains public sector information licensed under the Open Government Licence v3.0.
Buying a Tool Is Not Building a Capability
The uncomfortable truth in all this is that you can’t just buy AI as a single product and switch it on. It calls for a change in how people work, and that change needs support.
Naturally, that’s expensive and slow if you approach it as a pile of separate tools, each with its own login, quirks, and learning curve. No wonder so many businesses train nobody. If they’ve taken this approach, the mountain will simply look too high.
There’s a smarter route. Rather than bolting standalone AI tools onto the side of a business, the AI can live inside the systems people already use every day. When it sits inside your finance, sales, and stock system, the learning curve shrinks dramatically. Your team doesn’t need to learn a new tool, because AI quietly does more of the work inside the system they already know.
This is exactly how Copilot in Business Central works. It’s built into the platform your finance and operations teams already use, so there’s no separate app to master. It drafts and matches purchase invoices, answers plain-English questions about your own data, and handles routine tasks in the background. If you’ve ever wondered what the built-in AI actually does, the short answer is everyday work.
The training burden all but disappears, because the AI is just a better version of a thing staff already do.
Closing the Gap Without the Headache
You don’t have to panic about the 46% figure. It’s definitely a reason to act though. Businesses that keep treating AI as a set of gadgets to buy will keep struggling to see a return. That’s because the hard part was never the buying.
Those that build AI into their core systems, with proper support behind it, will pull ahead. Everyone else will still be writing training plans they never deliver.
We help UK businesses do exactly that. That means putting Microsoft Dynamics 365 Business Central at the centre of the business, with AI built in and senior consultants on hand. The aim is a team that can actually use it. No standalone tools gathering dust, no shadow AI creeping in around the edges, just capability that sticks.
Skill Up, Don’t Just Sign Up
If your business has bought into AI but not built the skills to use it, you’re far from alone, and there’s a better way forward. Let us show you how Business Central puts AI where your team can actually use it, with the support to match. Speak to the team about a system your people will get real value from.
Frequently Asked Questions
How Many UK Businesses Provide AI Training to Staff?
According to the ONS Business Insights and Conditions Survey, 46% of UK businesses that use AI haven’t given staff any AI training at all. Only around 11% of AI-using businesses with 10 or more employees report that more than half of their workforce has received AI-related training. So, meaningful, widespread training remains rare.
Why Does AI Training Matter So Much?
Because untrained AI is usually wasted AI. Tools that nobody knows how to use either go unused or get used without proper oversight, which raises risks around data security and accuracy. The most cited barrier businesses report is a lack of expertise, ahead of cost.
Which Businesses Are Least Likely to Train Staff on AI?
Smaller firms. Around 49% of AI-using businesses with fewer than ten employees have trained nobody, compared with 13% of firms with 250 or more staff. Manufacturing is also notably behind, with half of AI-using manufacturers having provided no AI training.
What Is Shadow AI?
Shadow AI is when employees use AI tools without employer approval, guidance, or oversight. It often fills the gap left by a lack of formal training, and it can lead to data leaks, confidentiality breaches, and compliance problems. The ONS finds 55% of employees use AI for work, against 35% of businesses reporting adoption, so much AI use already sits outside employer control.
How Can We Reduce the Training Burden of Adopting AI?
Put the AI inside systems your team already uses rather than adding separate tools. When AI is built into your core finance and operations platform, such as Business Central with Copilot, staff aren’t learning a new tool. The training burden shrinks and adoption sticks.
Where Does This Data Come From?
The figures are from the ONS Business Insights and Conditions Survey, Wave 159 (June 2026). They are supported by the ONS article “Artificial intelligence in UK businesses: 2023 to 2026”. Training figures relate to businesses that use AI or are not sure whether they do. You can read our full analysis in our UK AI adoption statistics study.
