22nd August 2026
There is a tendency whenever artificial intelligence is discussed to imagine that a revolution is already happening.
Jobs are supposedly disappearing, businesses are being transformed and computers are increasingly doing the work that people once did.
The latest figures from the Office for National Statistics suggest something rather more complicated.
British businesses are adopting AI at remarkable speed, but most are still only scratching the surface of what the technology can do.
That may actually be the most important finding in the ONS report.
The proportion of UK businesses with at least 10 employees reporting that they use at least one form of artificial intelligence has risen from around 12% in late 2023 to 35% by June 2026.
In less than three years, the proportion has almost tripled.
That is a remarkable rate of adoption.
But there is a catch.
For most businesses, AI has not yet become deeply embedded in the organisation.
The AI revolution is still surprisingly shallow
The ONS has found that the average number of different AI technologies being used by businesses has increased only from about 1.4 to 1.6 since 2023.
That suggests that although many more businesses are experimenting with AI, relatively few have transformed the way their entire organisation operates.
Indeed, among businesses using AI, only 10% describe their use as extensive.
Only around 15% say that more than half of their employees use AI as part of their daily work.
So the reality is rather different from the image of entire companies being run by artificial intelligence.
For most businesses, AI is still a tool being introduced here and there.
Perhaps someone is using it to write a first draft of a document.
Another employee may be using it to analyse information.
A marketing department may be generating images.
A business owner may be using a chatbot to help with research or administration.
The revolution is happening, but in many companies it is happening one task at a time.
The big companies are moving faster
Size matters.
The ONS found that 49% of businesses employing 250 or more people reported using AI, compared with 28% of businesses with fewer than 10 employees.
That makes intuitive sense.
Large companies have more money to invest, more data to work with and more employees whose jobs can potentially be reorganised around AI.
They also have dedicated IT departments and managers who can experiment with new technology.
For a small business, the situation is very different.
The owner may be the managing director, salesperson, accountant and IT department all at once.
There may be little time to investigate whether an AI system can actually save money.
And there may be legitimate concerns about putting commercially sensitive information into an AI system.
That creates a potential divide between large and small businesses.
But small businesses may actually have an advantage
There is another side to this.
A huge company may spend millions integrating AI into its systems.
A small company can sometimes simply start using it tomorrow.
That makes AI rather different from many previous technological revolutions.
A small business doesn't necessarily need to buy a factory full of expensive machinery.
It may need a computer, an internet connection and access to an AI service.
That could be particularly significant for Scotland's small-business economy.
A rural business in Caithness can potentially access many of the same AI tools as a company operating in Edinburgh, London or Manchester.
Geography becomes less important when the technology is delivered through the internet.
That does not mean the digital divide has disappeared, but it does create an unusual opportunity for smaller and more remote businesses.
What are businesses actually using AI for?
The answer from the ONS is revealing.
The most common purpose is improving existing business operations.
More than three-fifths of larger businesses using AI report this as a purpose.
In other words, businesses are generally not using AI to reinvent themselves.
They are using it to make what they already do more efficiently.
That could mean processing information more quickly, helping with administration, generating content, analysing data or improving existing processes.
It is a much less dramatic form of technological revolution.
But it could ultimately be more important.
If AI saves somebody in a small business an hour a day, that may not make a newspaper headline.
Multiply that across millions of employees and the potential productivity effect becomes much more significant.
And here comes the really interesting statistic
The ONS found that 55% of employees report using AI for work or education, compared with only around 35% of businesses reporting that they use AI.
That tells us something important.
AI may already be spreading through British workplaces faster than companies themselves realise.
Employees can experiment with AI independently.
A worker may use an AI assistant to improve an email, summarise a document, research a subject or help with a spreadsheet without the business having formally introduced an organisation-wide AI system.
That creates both an opportunity and a problem.
The opportunity is obvious.
Employees are finding their own ways to become more productive.
The problem is that businesses may not know exactly what their employees are doing.
The unofficial AI revolution
This could become one of the most important aspects of AI adoption.
Imagine an employee who discovers that an AI system can perform a task that normally takes them two hours in ten minutes.
They start using it.
The business becomes more productive.
But nobody formally introduces an AI policy.
Nobody measures the time saved.
Nobody knows precisely what information is being put into the system.
The technology has entered the company almost invisibly.
That may explain why the employee figures are so much higher than the business figures.
The AI revolution may therefore be occurring from the bottom up rather than from the boardroom down.
The threat to jobs is also more complicated than the headlines suggest
The ONS found that AI use is already affecting particular types of work.
More than half of businesses using visual-content AI technologies report effects on creative or design roles, while a similar proportion using AI image-processing technologies report effects on administrative or clerical roles.
But the ONS has not found widespread changes in overall workforce headcounts as a result of AI.
That is an important distinction.
AI can change what somebody does without eliminating their job.
A graphic designer might use AI to generate initial ideas rather than spending hours creating them from scratch.
An administrator might use AI to process routine information while spending more time dealing with customers.
A journalist might use AI for research and background information while retaining responsibility for deciding what should actually be published.
The job changes before the job disappears.
That may be the pattern we see most often.
Britain may be entering an AI productivity experiment
This is perhaps where the report becomes economically significant.
Britain has struggled with weak productivity growth for years.
If AI can allow existing workers to produce more without proportionately increasing costs, it could provide a much-needed boost.
There are already signs that UK productivity growth has begun to improve, although economists remain cautious about attributing the improvement directly to AI.
The problem is that productivity improvements don't automatically translate into higher wages.
A company could use AI to produce the same output with fewer hours of labour.
It could use the savings to increase profits.
It could reduce prices.
Or it could invest the money elsewhere.
What happens to workers will depend on how businesses choose to use the productivity gains.
The biggest obstacle may not be technology
The ONS found that 41% of businesses with 10 or more employees reported experiencing no barriers to AI adoption.
Among those that did encounter difficulties, lack of expertise was one of the most frequently reported problems.
Cost was another significant barrier.
That tells us something important.
The question for many businesses is no longer:
“Does AI exist?”
It is:
“Do we know how to use it properly?”
That is a very different problem.
A business can buy access to an AI system in minutes.
Finding somebody who understands how to integrate it into the company's processes, protect confidential information, check its output and make sure employees use it effectively is much harder.
The danger of becoming an AI follower
There is also a competitive issue.
If one company in an industry uses AI to reduce its administrative costs by 20% while its competitors do nothing, it may eventually gain a significant advantage.
That could force competitors to follow.
This is how technological adoption can suddenly accelerate.
For several years, nothing much happens.
Then enough businesses adopt the technology that those remaining outside it begin to feel they are falling behind.
The ONS figures may therefore represent an early stage of a much bigger change.
Thirty-five per cent adoption today does not necessarily mean 35% tomorrow.
What does this mean for Scotland?
There is a particularly interesting question here for Scotland.
AI potentially offers smaller and more remote businesses something they have historically struggled to obtain: access to sophisticated capabilities without having to be physically close to a major city.
A business in Caithness can use AI for marketing, customer communications, research, administration, translation, data analysis, software development and increasingly sophisticated creative work.
That doesn't remove the disadvantages of distance.
But it can reduce some of them.
A small Highland business cannot employ a specialist in every field.
AI may allow one person to perform tasks that previously required several different specialists.
That could become a genuine competitive advantage.
But there is a warning
AI isn't automatically productive.
A badly implemented AI system can create more work rather than less.
It can produce inaccurate information.
It can create security and privacy risks.
It can encourage businesses to publish material that nobody has properly checked.
And it can make employees less willing to develop their own skills if they simply accept whatever the machine produces.
The ONS figures actually provide some reassurance on this point.
The technology is spreading rapidly, but the relatively shallow depth of adoption suggests that most British businesses are still experimenting rather than handing control of important operations entirely to machines.
Perhaps that is exactly where businesses should be.
The AI revolution may be quieter than we imagined
The most interesting conclusion from the ONS report is that the AI revolution may not arrive with a dramatic moment when computers suddenly replace people.
It may happen much more quietly.
An employee saves 30 minutes.
A business processes an invoice automatically.
A marketing department produces its first draft in an hour rather than a day.
A small company analyses information that previously required an expensive consultant.
A shop owner uses AI to understand which products are selling.
A rural business gains access to expertise it could never afford to employ.
None of those changes is revolutionary by itself.
But put millions of them together and the economic consequences could be enormous.
The ONS figures show that Britain has already moved a long way down that road.
AI use among businesses with 10 or more employees has almost tripled since 2023.
Yet only a minority of those businesses are using it extensively.
That leaves us with a fascinating possibility.
The biggest impact of AI on Britain's economy may not have happened yet.
The first phase has been about discovering what AI can do.
The next phase could be about discovering what happens when businesses genuinely reorganise themselves around it.
And that is when the argument over jobs, productivity, wages and competitiveness becomes much more serious.
The machines may already be in the workplace.
The real AI revolution may begin when businesses learn how to use them properly.
Source
The ONS report at
https://www.ons.gov.uk/businessindustryandtrade/business/businessservices/articles/artificialintelligenceinukbusinesses/2023to2026