UK firms pour into AI despite weak returns, study finds

UK businesses are investing heavily in artificial intelligence despite weak returns, according to research commissioned by Expereo. The study found that network shortcomings are one reason many deployments have failed to meet expectations.

IDC's research found that 55% of UK organisations plan to prioritise AI or machine learning for effort or financial investment over the next 12 months. Yet only 15% said their AI implementations had exceeded expectations, and just 3% said they had significantly exceeded them.

Two-thirds of UK organisations said they were investing in AI because of its potential, while 15% said they were investing aggressively with little evaluation because they feared being left behind. The findings point to a market where spending is rising faster than businesses can assess whether projects are delivering a return.

Several factors were cited for weaker-than-expected outcomes. Nearly half of respondents, 48%, blamed inadequate or poor-quality training data, while 45% pointed to higher-than-expected costs or a failure to achieve return on investment. A further 42% said AI was not performing as well as expected.

Network performance also emerged as a constraint. Some 24% of UK businesses said inadequate network or connectivity performance was a reason their AI implementations had failed to meet expectations over the past year.

Infrastructure gap

The survey suggests many organisations now see network upgrades as a necessary part of broader AI spending. More than half, 54%, said they need more flexible and scalable networks to operate in an AI-driven environment, while 57% said they need greater resilience and reliability to maximise uptime as AI workloads increase.

That matters because AI systems often depend on moving large amounts of data between users, cloud platforms and internal systems. If networks are unreliable or lack capacity, businesses can face delays, higher operating costs and weaker performance from the tools they have bought.

Expereo, a provider of managed network-as-a-service products, said the figures show a widening gap between boardroom expectations and operational readiness. As AI programmes move from pilots to wider use, network decisions are becoming more closely tied to investment outcomes, it said.

Ben Elms, Chief Executive Officer at Expereo, said: "Every UK organisation we speak to is investing in AI, yet the data shows a clear gap opening up between AI ambition and AI outcomes. More often than not, that gap comes down to the network underneath. AI only delivers on its promise when the infrastructure carrying it is built to support it. Without resilient, scalable, cloud-optimised networks, even the most well-funded AI programmes will struggle to deliver ROI. Getting the network right is no longer an IT decision; it is one of the most important conversations happening in the boardroom today to help fulfil AI ambition."

Risk concerns

The findings also show that concerns are spreading beyond technical performance into governance and financial control. More than half of UK technology leaders, 56%, said the creation of new security risks was a significant potential future threat linked to their organisation's use of AI.

Cost oversight was another concern. A third of respondents, 33%, said they were worried about losing track of AI-related costs and return on investment once the technology becomes embedded across the business.

These concerns reflect a broader pattern in the UK market, where pressure to adopt AI has intensified even as many organisations are still trying to build the systems needed to support it. Businesses may be moving quickly to avoid missing out on competitive gains, but the research indicates many are doing so without the infrastructure discipline needed to make those investments pay off.

For network suppliers, telecoms groups and technology service providers, the data also highlights a commercial opening. If companies continue to prioritise AI spending while struggling with reliability, scale and connectivity, attention is likely to shift towards the less visible infrastructure supporting those projects.

The IDC study covered attitudes to investment priorities, performance and risk, and its UK findings suggest the debate over AI returns is moving away from headline spending towards practical questions about delivery. With only 15% of organisations saying AI projects are exceeding expectations, pressure is likely to grow on boards to show that the systems underneath those tools are fit for purpose.