Press
UnlikelyAI
99% of senior decision-makers are already using AI. Three years into mainstream adoption, only 22% report significant returns. Techround picked up our AI Trust Ceiling research alongside exactly that question, if businesses are already using AI, why isn’t it working?
The readiness gap
The UK government’s ambition to lead the G7 in AI adoption has an engineering problem underneath it. Our data shows 99% of senior decision-makers are already using AI, adoption is not the constraint. What follows adoption is that 87% say they trust AI in principle, 99% verify its outputs before acting on them. That gap is what happens when systems cannot show their reasoning and no amount of faster adoption closes it.
Faster adoption of broken tools is not a growth strategy
UK businesses are not short of AI appetite. They are short of AI they can rely on for work that matters. The dominant tools in the market of large language models were designed to produce outputs that sound right, and are statistically correct. For content generation or summarisation, the distinction is manageable. For the high-stakes decisions where AI would deliver the largest productivity gains in risk, compliance, and strategy it is the entire problem. Faster adoption of tools with that limitation does not unlock value, it scales the verification burden.
“”The next generation of AI infrastructure must look beyond the simple attraction of AI ease and focus on the integrity of AI decisions. The wrong decisions, if made due to rushed AI adoption, could impact millions of people and cost the economy billions.” — Stuart Harvey, as quoted in Techround”
What building trustworthy AI infrastructure means
Our research found that 30% of leaders would increase AI investment immediately if accuracy and explainability were guaranteed. The investment is ready but the architecture is not.
Technology that produces outputs that can be verified, audited, and explained at each step is that condition those 30% are waiting for.
Speed without trustworthiness is not a competitive advantage. It is a liability that compounds every time an unexplainable AI decision lands in a high-stakes context and those are exactly the contexts the UK’s regulated industries cannot afford to get wrong.
Not ready to speak to us? Start here.
See the sector by sector breakdown
