It will come as no surprise that organisations across the UK have spent the last two years pouring significant money into artificial intelligence. The UK AI sector alone attracted a record £2.9 billion in investment in 2024. Pilots have been launched, tools have been procured, and technology teams have been expanded.
Yet for many, the returns remain stubbornly elusive. The gap between expectation and reality is widening, and the question being asked in boardrooms with increasing urgency is why.
The answer, in most cases, is not the technology. It is the way organisations have chosen to deploy it. Until businesses stop treating AI as a technology problem and start treating it as a business transformation, that gap will not close.
AI is not an IT problem
The instinct to hand AI to the technology function makes sense in theory. It involves data infrastructure, integration, security and systems architecture. Now its not surprising that these read like an IT bingo card as they are clearly all technical disciplines, and they matter. But treating AI as an IT project is where many organisations make their first and most costly mistake, because it places accountability for outcomes in the wrong part of the organisation.
Technology can enable AI, but it does not create value on its own. Value is created when AI is embedded into how the business actually operates: how decisions get made, how workflows are structured, how people are managed, how outcomes are measured and held to account. These are not technology questions. They are business questions, and they require business ownership, if you are asking your technology leader to answer these questions then should already know you are not going to see the value from AI within your business.
Leading organisations recognise this, especially learning from their previous digital transformation journeys and are moving AI out of central technology teams and into the functions where value is actually created –operations, finance, customer experience, supply chain. Technology still plays a critical role, providing the architecture, controls and integration layer. But the initiative, the workflow redesign and the accountability for outcomes sit with the business. That distinction is not a subtle one, it changes everything about how AI programmes get prioritised, resourced and sustained.
The slap-on AI problem
Misplacing ownership has a direct consequence, and it shows up in a pattern that has become familiar across industries. A company identifies a process that feels like a good candidate for AI. A tool or platform is selected, a pilot is run, and early results look promising. Then the pilot ends, broader rollout stalls, and the anticipated value fails to materialise. What went wrong?
In most cases, the tool was layered onto an existing process rather than used as a reason to rethink it. This is the ‘slap-on’ AI problem – and it is more widespread than many organisations would care to admit. Bolting AI onto fragmented data, unclear decision rights, inconsistent processes and limited change capability does not produce transformation. It produces a more expensive version of the same issue at best and an increasingly likely more complex organisation
The problem is compounded when organisations treat AI as something that can be piloted in isolation and then scaled by default – as the assumption is that this is ‘just another’ digital capability. Real value rarely works that way. It requires workflows to be redesigned end to end, not just touch the edges. You need clarity about who owns decisions and what success actually looks like alongside the kind of change management capabilities which most technology-led AI programmes simply do not prioritise. Without business ownership driving that redesign, slapping a tool onto a broken process will always produce the same result.
From experimentation to sustained value
Fixing this requires more than a change in mindset. It demands a different set of operational commitments, and the organisations making genuine progress tend to share the same characteristics. They are not necessarily the ones running the most experiments as many business leaders will suggest is a critical success factor.
Leadership is the most important factor. AI transformation is not a technology programme with a business sponsor – it is a business transformation that happens to be enabled by technology. Sound familiar?, well it should, the businesses who are successful sooner in driving their AI transformation will likely originate from those organisations that realised sustainable change from their previous Digital transformation efforts. This is why that framing has to come from the top, and it has to be sustained. Organisations where senior business leaders are genuinely accountable for AI outcomes consistently outperform those where accountability rests with the technology function alone.
Funding models matter too. Many organisations are still committing large budgets to AI based on broad strategic ambition rather than specific outcomes. Smaller, targeted investments tied to clear business problems and measurable results tend to deliver far better returns and build organisational confidence in a way that large, diffuse programmes rarely do. Think about a portfolio approach where you are able to consistently make persist, pivot and kill approaches to initiatives, noting that a handful of initiatives will likely scale, but their success aides in funding the next wave of identified opportunities.
Ultimately, what gets measured gets managed. If AI performance is being tracked through technology metrics – adoption rates, system uptime – the organisation is measuring the wrong things, because while operationally may be positive there is zero line of sight to business impact. The metrics that matter are business metrics like cost reduction, decision speed, revenue impact, order fall out and risk reduction. Without these, it is almost impossible to separate genuine progress from activity, and almost impossible to make the case for the next investment.
The hard changes are the ones that count
There is no shortage of enthusiasm for AI in the UK, much like anywhere else in the world. What remains in shorter supply is the willingness to make the structural changes that give it the best chance of delivering stellar results for UK businesses.
Redesigning workflows, moving ownership into the business, changing how investment is allocated and how performance is measured. None of these are straightforward. But they are precisely what separates organisations that are realising value from those that are still waiting for it.
AI will not deliver on its promise by being treated as ‘another digital or IT initiative’. It will deliver when businesses are willing to change how they actually run and ensure AI’s success is owned where the value is to be realised.
