Six in ten businesses say an AI skills gap is holding them back from adopting artificial intelligence, making it the most commonly cited barrier to adoption, ahead of cost, unclear regulation, and ethical concerns, according to new data from the Department for Science, Innovation and Technology (DSIT).
The research, described as the most comprehensive study of its kind and commissioned to support the UK government’s AI Opportunities Action Plan, found that only 1 in 6 UK businesses (16%) currently use AI, with 80% having no active plans to adopt it. Among businesses that have already tried to roll out AI more widely, limited skills and expertise remains the single biggest obstacle, cited by 54% of existing AI adopters as a hindrance to scaling further.
Nasstar, one of the UK’s leading managed IT providers, says the findings lay bare a structural challenge facing British businesses: the gap between awareness of AI’s potential and the internal capability to act on it.
That gap is increasingly visible in real-time search data, too. Google Trends analysis conducted by Nasstar shows that UK search interest in “how to use AI” increased 90-fold between January 2022 and early 2026, reaching an all-time peak in December 2025. Searches for “AI use cases,” a signal that people are actively trying to understand what AI could do for them specifically, were virtually non-existent before 2024 and hit a record high in January 2026. Together, the data suggests that awareness of AI is no longer the issue: it is the practical knowledge of what to do next that remains elusive.
Sean Morris, Chief Technology Officer at Nasstar, commented:
“The government’s research confirms what we hear from businesses every day: the challenge isn’t a lack of ambition around AI, it’s a lack of the expertise needed to get started safely and at scale. Most organisations know they should be doing more with AI, but they’re not sure where to begin, and they don’t have the people internally to figure it out.
“The good news is that the tools most UK businesses need are already sitting inside Microsoft 365, in the applications they use every day. The barrier isn’t access to technology. It’s knowing how to unlock it.”
The skills gap in numbers
The DSIT data paints a clear picture of where the skills deficit bites hardest. Sixty per cent of all UK businesses cite limited AI skills, expertise, or knowledge as a factor preventing adoption, the second most commonly cited barrier overall, behind only a perceived lack of need. Among businesses already using AI, 54% say limited skills are actively hindering them from scaling further across their organisation.
The figures are starker still among those planning future adoption. Sixty-eight per cent of that group cite skills as a barrier, higher than any other obstacle, including cost (23%) and unclear regulation (28%). Only 1 in 3 businesses planning to adopt AI feel ready to implement it, compared with just over half (54%) of existing adopters who feel prepared to scale further.
The divide between larger and smaller businesses is also pronounced. Large businesses adopt AI at more than double the rate of micro businesses (36% versus 14%), a gap driven in large part by the resources available to invest in AI skills and expertise.
The research also reveals that most businesses that do adopt AI rely entirely on off-the-shelf solutions rather than in-house development, with 71% of natural language processing adopters purchasing ready-to-use external tools. Qualitative interviews within the study repeatedly surfaced the same explanation: organisations simply do not have the technical expertise to build or configure AI systems themselves.
Nasstar’s analysis of Google Trends data reinforces the picture. UK search volumes for “learn AI” rose by almost 2,000% between January 2022 and January 2026, with similar spikes recorded for “AI use cases” and “how to use AI”. The search data points to a workforce that knows it needs to engage with AI but is still working out how and where to start.
The productivity opportunity, and why it remains out of reach for most
The case for action is not merely about keeping pace. Three-quarters (75%) of businesses that have adopted AI report improved workforce productivity, and over half (57%) say they have developed new or improved processes as a result. Adoption is also associated with cost reductions (34%), improved staff retention and wellbeing (19%), and entry into new markets (18%).
The most commonly cited reason for wanting to adopt AI, among both current and prospective adopters, was to increase efficiency or productivity, cited by 65%. Yet the majority of UK businesses remain on the outside looking in, with the skills barrier the principal reason they cannot cross the threshold.
Separate research from Microsoft, based on a Censuswide survey of 1,000 UK senior decision-makers published in January 2026, found that 84% of organisations are now deploying AI for competitive advantage, up from 40% in 2025. Nearly a quarter of leaders said they were worried their competitors were adopting AI faster than their own organisation. For businesses that lack the internal expertise to keep pace, that anxiety is unlikely to ease without external support.
Ash Ward, Comms Capability Lead at Nasstar, added:
“There’s a risk here that isn’t being called out enough. AI won’t level the playing field; it’ll widen the gap between organisations that know how to use it and those that don’t. The data already shows the issue. The skills gap isn’t something on the horizon. It’s here now, and it’s hitting smaller businesses harder than most.
“The technology itself isn’t the blocker anymore. Tools like Microsoft Copilot are already built into Word, Excel, Teams, and Outlook, platforms people use every day. What’s missing is the structure around it. Clear use cases, a plan for adoption, and the right guidance to make it actually work for us.
“For businesses unsure where to start, Microsoft has made a lot of this accessible. There are hundreds of free Copilot training modules available through Microsoft Learn, many aligned to specific roles and industries. The capability is there. It comes down to whether organisations choose to build it.”
