AI investment is rising but UK organisations are struggling to see meaningful impact from their efforts.
According to the 2026 AI Readiness Report by Southampton-based digital product studio Rareloop, growing budgets are being hampered by outdated systems and disconnected data, with nearly 80 per cent of organisations surveyed planning to increase AI spending this year.
The report, based on responses from more than 100 senior leaders across SMEs, scale-ups, and enterprise organisations, found that infrastructure, rather than budget, is now the primary barrier to AI success.
READ MORE: New electric vehicle brand launches in Southampton
Joe Lambert, founder of Rareloop, said: “AI isn’t failing because organisations lack ambition.
“It’s failing because their data is locked away, their systems don’t talk, and their architecture wasn’t built for integration.
“Readiness, not tooling, is now the constraint on value.”
Organisations reporting the highest levels of AI success were often those investing under £5,000 annually, while larger organisations with much bigger budgets reported similar or lower satisfaction due to difficulties integrating AI with legacy systems.
53 per cent of organisations surveyed rely mainly on off-the-shelf tools such as ChatGPT, Google Gemini, and Microsoft Copilot, while 45 per cent identified legacy systems and poor data access as the biggest barriers.
The report suggests AI is acting as a “stress test” for existing technology, revealing long-standing issues like technical debt, fragmented systems, and inaccessible data.
Organisations often use AI to compensate for inefficient processes or poor user experience, instead of addressing core infrastructure problems.
READ MORE: Business Voice – Investing in innovation
Those seeing the most value from AI are investing in system integration, data accessibility, and scalable architecture before introducing AI capabilities.
The research also highlighted a shift from experimentation toward integration.
While AI adoption is widespread, many organisations remain in an experimentation phase, using standalone tools instead of embedding AI into core workflows and products.
The report concludes that long-term value will not come from adding more tools, but from improving system connectivity and data flow across the organisation.