HRMS News logo

AI skills shortage holds back 42% of UK businesses

AI skills shortage holds back 42% of UK businesses

New research from Turbotic has found that 42% of UK businesses identify a lack of AI skills and expertise as the biggest challenge when trying to scale AI initiatives, according to HR News. The shortage of expertise ranked as the second most significant obstacle overall, behind integration with existing systems at 47% and ahead of securing budget at 36%.

The study surveyed more than 1,000 senior decision-makers involved in setting and assessing AI projects. It found that 34% of all respondents said insufficient internal AI skills was the single biggest reason projects struggle to progress beyond the pilot stage. Just 16% reported successfully scaling AI across their organisation, while 37% said they can build AI pilots but struggle to scale them.

Why the skills gap is structural, not temporary

Theodore Bergqvist, CEO and Co-founder of Turbotic, said Britain cannot educate its way out of the AI skills shortage quickly enough. The technology is developing faster than conventional training programmes can keep pace, and businesses are discovering that knowing how to use AI is very different from knowing how to build systems that actually work.

The research points to a deeper problem than simple headcount. Organisations can hire or train people who can operate AI tools, but the shortage sits at the level of system design, integration and reliability. Bergqvist noted that the people who understand how different AI models interact, how to integrate them into existing operations and how to make them reliable at scale are in desperately short supply. That expertise is scarce even among the world’s biggest technology companies, never mind smaller firms with fewer resources or those in sectors with more complex needs.

What this means for HR and payroll teams

For HR and payroll leaders, the findings carry practical implications. Payroll systems increasingly incorporate AI-driven anomaly detection, error flagging and compliance monitoring. HR teams are piloting AI for recruitment screening, absence pattern analysis and workforce planning. The Turbotic data suggests many of these initiatives will stall between pilot and production.

The sector breakdown is telling. Financial services reported the highest level of skills and expertise as an obstacle to scaling AI at 45%, followed by healthcare at 43%, distribution and logistics at 42%, and education at 40%. Payroll and HR functions within financial services and healthcare face particular pressure, because those sectors are deploying AI in areas such as fraud detection, risk assessment and compliance, where technical expertise needs to be matched with an understanding of operational processes and regulatory requirements.

The size split matters too. Some 46% of businesses with fewer than 1,000 employees cited skills and expertise as a challenge when scaling AI, compared with 38% of those with 1,000 or more employees. Smaller employers are less likely to have dedicated AI engineering roles, which means HR and payroll leaders in those organisations may be asked to manage AI projects without specialist support.

AI skills shortage holds back 42% of UK businesses (illustration)
Photo: The Skills Gap – Concept Illustration by Digits.co.uk Images, CC BY 2.0

The pilot trap and how to avoid it

The finding that 37% of organisations can build AI pilots but struggle to scale them points to a common failure pattern. Pilots are often run by a small team with external support, using clean data and limited integration requirements. Scaling requires the AI to work with existing systems, handle edge cases and meet governance standards.

The research identified integration with existing systems as the top obstacle at 47%. For payroll specifically, this means AI tools must connect with HRIS platforms, time and attendance systems, benefits administration and general ledger software. A pilot that works on a spreadsheet extract will not necessarily work when connected to live payroll data.

HR and payroll leaders should ask three questions before approving any AI pilot. First, who will own the integration work if the pilot succeeds? Second, does the organisation have anyone who understands how the AI model behaves when data quality varies? Third, what governance and compliance checks will apply when the tool moves into production? The Turbotic study’s wider findings highlighted concerns around governance and compliance, and fears employees are feeding sensitive company data into unauthorised AI tools.

The immigration and training tension

Bergqvist argued that Britain needs to be much more aggressive about attracting experienced AI specialists from overseas. Other countries will be competing for exactly the same people, and businesses cannot afford to wait years for domestic training programmes to catch up. He said the UK needs to bring experienced talent into the country now, while investing in the skills that will allow British businesses to build and manage these systems themselves in the longer term.

The UK government announced a £700 million AI and AI security investment package during the Budget 2026 statement. Demand for UK IT contractors reached a 44-month high in September 2026, according to the HR News report. For HR leaders, this means recruitment for AI-adjacent roles will remain competitive and expensive. Payroll teams should expect contractor costs for AI integration work to remain elevated.

Obstacle to scaling AI Percentage of UK businesses
Integration with existing systems 47%
Lack of skills and expertise 42%
Securing budget 36%

The practical response for most organisations will be a mix of targeted hiring, selective outsourcing and internal upskilling focused on operational AI literacy rather than deep technical training. HR teams should plan for longer lead times on AI-related roles and consider whether some integration work can be contracted while internal capability develops.

Sources: HR News

Photo: Streamax AI technology applications in the railway industry (32642645637) by International Railway Summit, CC BY-SA 2.0

administrator

Related Articles