UK organisations have never spent more on technology, and their staff have never complained about it more. Every office has a version of the same story: the new system nobody asked for, the software that turned a ten-minute job into a thirty-minute one, the rollout that arrived by email with a login and no explanation.
The pattern repeats because the thinking behind it repeats. Technology gets chosen for what it promises on paper, then dropped onto the people expected to use it. And when that thinking goes unchallenged for long enough, it can end somewhere far worse than a clunky rollout, as one British institution found out in the most public way possible.
When the System Wins Every Argument
The Post Office spent two decades prosecuting sub-postmasters because its Horizon accounting software said money was missing. Hundreds of people lost their livelihoods, their reputations and in some cases their liberty, while the staff reporting faults were treated as the problem. It took a public inquiry and years of campaigning to establish what those staff had said all along: the system was wrong.
Most organisations will never produce a scandal on that scale, but plenty run on the same quiet assumption. When the software and the person disagree, the software wins. Staff who find a tool unusable get told to follow the process, so they build workarounds instead, and the workarounds become the real system. IT then gets blamed for problems that were baked in at the point of purchase.
The cost rarely shows up on any invoice. It shows up in lost hours, in good people leaving, and in decisions made on data everyone quietly knows is unreliable.
AI Is Repeating the Pattern at Speed
Almost a third of UK businesses now use AI in some form, according to ONS figures, up from 9% in September 2023. The same data shows how shallow much of it runs, with the average adopter using just 1.6 AI technologies. Plenty of firms have bought the licence, announced it in a Monday meeting and moved on.
That’s the old mistake at a faster pace. AI genuinely will save time on repetitive work, from drafting documents to summarising long threads, and it will free people for work that needs judgement. But a tool dropped on an untrained team fails the way every tool before it failed, except this one can also leak client data when staff paste it into whichever free chatbot they found first.
The Alternative Starts With a Different Question
Organisations that get results from technology tend to open with a different question. Not what can this system do, but how do our people actually work, and what gets in their way. It’s the principle behind companies offering high-quality, people-focused IT solutions, where systems and support get shaped around staff, not the other way round.
Worked through properly, that question changes most technology decisions. It means asking the team where their week disappears before buying anything, training people until they’re confident instead of sending a single demo video, and treating a complaint about a tool as evidence rather than resistance. The principle holds whether the project is a cloud migration or a first AI use case.
It also means taking staff seriously when they say something isn’t working. The Horizon scandal ran for twenty years on the refusal to do exactly that.
Four Questions to Ask Before Any Rollout
- Whose working week will this improve, and can they name the specific task? If nobody can, the purchase is serving the buyer, not the users.
- Who will train people, and for how long after launch? Confidence comes from repeated use on real work, not from a launch email.
- What data can go into it, and who decided? Unwritten rules will be broken by accident within a month.
- What happens when it breaks, and who will staff call? The answer tells you whether people or systems come first in practice.
Where the Divide Opens From Here
Regulation will force some of this discipline anyway. The EU AI Act’s obligations are arriving in stages through 2026 and 2027, and UK firms selling into Europe will be caught by parts of it, while the ICO already expects personal data in AI systems to be handled with the same care as anywhere else. Organisations that have built good habits around people will absorb those rules without much drama.
The bigger divide, though, won’t be drawn by regulators. Over the next few years, AI and new-tool adoption figures will keep climbing across the board, and the results will cluster around the organisations that trained, listened and shaped technology around their staff. Everyone else will keep buying tools, and keep wondering why nothing changes.
David Prior
David Prior is the editor of Today News, responsible for the overall editorial strategy. He is an NCTJ-qualified journalist with over 20 years’ experience, and is also editor of the award-winning hyperlocal news title Altrincham Today. His LinkedIn profile is here.
