I have been building software for 27 years. I have worked through the dot-com boom, the cloud migration wave, the mobile revolution, and now the AI explosion. If you asked me to distil everything I have learned into one sentence, it would be this: buying a tool is the easy part. Using it well is where businesses actually win or lose.
That sounds obvious. It is obvious. And yet almost every business I have ever worked with has fallen into the same trap. They buy the thing, they announce the thing, they put the thing on a slide for the board. Then six months later nobody is using the thing and nobody can explain why.
AI is repeating this pattern right now, at scale, across the UK.
The Numbers Tell the Story
The British Chambers of Commerce, in partnership with Atos, published their latest research in early 2026. AI adoption among UK SMEs has climbed to 54%, up from 35% in 2025 and just 23% in 2023. That is a genuine, rapid increase. More than half of UK SMEs are now actively using AI in some form.
But here is the number that should make you pause. Government research published in June 2026 found that only 21% of UK workers feel confident using AI at work.
54% adoption. 21% confidence.
That gap is where money gets wasted and opportunities get missed. It is the gap between having a tool and having a capability. And it is the same gap I have watched businesses fall into for nearly three decades.
The Pattern I Have Seen for 27 Years
In 1999, I watched companies buy enterprise content management systems that nobody ever used. In 2008, I watched them migrate to cloud platforms and then run them exactly like their old on-premise servers. In 2015, I watched them build mobile apps that got downloaded 200 times and then abandoned.
The technology was fine. The technology was usually excellent. The problem was always the same: somebody confused purchasing with capability.
Here is what happens. A decision-maker sees a demo, reads a case study, or attends a conference. They get excited. They buy the licence, the platform, the system. They tell the team “we now have X.” And then they move on to the next decision, assuming the job is done.
But the job has not started. The job is not the purchase. The job is the hundred small decisions that follow: how does this fit into our daily work? What does the old process look like now? Who needs training? Who needs permission? What do we stop doing because the new tool handles it?
Those questions are unglamorous, time-consuming, and easy to skip. So most businesses skip them.
Why AI Makes This Worse
AI intensifies the purchase-versus-capability gap for three reasons.
First, AI tools change constantly. The platform you bought in January may have different features by July. This is not like buying an accounting package that does the same thing for five years. You are buying into a moving target, which means capability has to be continuously rebuilt.
Second, AI tools are open-ended. A CRM has a clear purpose. You put customer data in, you get pipeline reports out. AI tools can do dozens of things, which means without clear direction, people use them for nothing. Choice paralysis is real. I have seen teams with access to powerful AI tools who use them for nothing more than email drafting because nobody told them what else was possible.
Third, AI capability is deeply personal. Two people using the same tool can have wildly different outcomes. One person writes a prompt that saves them three hours a week. Another person types “write me a report” and gets generic mush back. The tool is identical. The capability is not.
What Capability Actually Looks Like
The BCC research found something interesting buried in the data. Firms currently deploying AI report a net productivity expectation of +71%. Firms planning to adopt but not yet doing so report +46%. Firms with no plans report -3%.
The firms actually using AI are seeing results. The firms still thinking about it are seeing less. And the firms ignoring it are going backwards.
But “deploying AI” in this context does not mean buying a licence. It means people are using the tools in their daily work, regularly, for real tasks. That is capability.
The government’s Skills for AI research programme, published in June 2026, backed this up. Their survey of 536 employers found that over 44% of organisations now use AI tools daily. But they also found that most staff learn through “trial and error, peer support, online videos or built-in tool prompts.” That can get you started, but it creates “uneven and risky practice.”
In other words, people are figuring it out on their own, inconsistently, with no shared standard for what good looks like. Some people in your organisation are probably brilliant with AI already. Others have never opened the tool you are paying for. The gap between your best user and your worst user is enormous, and you probably do not even know who is who.
Only 11% of UK businesses reported that more than half their workforce has received any AI-related training. Think about that. We are in the middle of the fastest technology shift in a generation, and 89% of businesses have not trained even half their staff.
The Lesson, Applied
So here is what 27 years has taught me about closing the gap between purchase and capability. It comes down to four things.
Pick one tool, not five. I have seen businesses buy three AI platforms simultaneously and use none of them well. Choose one. Use it for three months. Build fluency. Then decide if you need another.
Attach it to a specific task. Do not tell your team “we now have AI.” Tell them “we are using AI to draft our client proposals, and here is how.” Specific tasks create specific habits. Vague mandates create nothing.
Train in short bursts, not marathon sessions. The government’s research found that 30 to 90 minute modules work better than long courses for SMEs. This matches what I have seen. An hour of practical, task-specific training beats a full day of theory every time.
Make someone responsible. Not a “AI champion” with a badge and no authority. Someone whose actual job includes making sure the team uses the tools well, shares what works, and flags what does not. In a small business, that might be 30 minutes a week. But it needs a name attached to it.
The Honest Truth
The businesses that will benefit from AI over the next two years are not the ones with the biggest budgets or the fanciest tools. They are the ones that close the gap between buying and using.
That gap has existed in every technology wave I have lived through. It is not new. What is new is how fast AI is moving, which means the cost of not closing the gap is higher than it has ever been.
If you are a UK SME owner reading this and you have bought an AI tool but your team is not really using it, you are not alone. 54% of your peers are in the same boat. But being in good company does not make it a good strategy.
Close the gap. Pick a tool, pick a task, train your people, and make someone accountable. It is not complicated. It is just hard, unglamorous work. The kind of work that actually pays off.
I have been watching this pattern for 27 years. The businesses that do this work win. The ones that do not, eventually stop talking about the tool they bought and hope nobody asks what happened to it.
Do not be that business.