AI adoption challenges: what UK business leaders said about digital transformation and AI
Digital technology and artificial intelligence (AI) were the focus of a round table attended by key industry figures, including Perninder Dhadwar, CEO of Imobisoft. The discussion, organised by Insider Media, brought together a diverse group of fast-growing companies, funders, advisors and marketers. They explored how AI is transforming business operations, the opportunities it presents and the AI adoption challenges it brings.
The panel's main concerns were finding a real problem for AI to solve, cost, skills, data privacy and trust. We first published this summary of the discussion in June 2024. For anyone weighing up digital transformation and AI today, we have added UK government figures from 2025 and 2026. They show many of the same barriers the panel raised.
What are the biggest AI adoption challenges for UK businesses?
The most common AI adoption challenges are not knowing where AI would help, limited skills and a lack of tools. The Department for Science, Innovation and Technology (DSIT) asked 3,500 UK businesses what has held back their AI adoption. Its survey, carried out between February and May 2025 and published in January 2026, found these barriers to AI adoption:
| Barrier | Share of businesses |
|---|---|
| Have not identified a use for AI | 71% |
| Limited AI skills, expertise or knowledge | 60% |
| Lack of tools or platforms for developing AI models | 48% |
| Regulation is unclear or uncertain | 28% |
| Cost is too high | 23% |
| Ethical concerns | 22% |
| Too much data complexity | 21% |
Cost came only fifth. The Office for National Statistics (ONS) measures the question differently. In its June 2026 survey, 41% of businesses with 10 or more employees reported no barriers. Among the rest, a lack of expertise was one of the most common answers. Meanwhile, cost affected between about 7% and 14% of businesses, depending on size.
For more on how larger UK firms handle these barriers, see our whitepaper on AI adoption in UK enterprises.
Who discussed AI adoption challenges at the round table?
Insider Media's Growth through Digital & AI round table brought together investors, advisors, a law firm, a university and technology businesses:
- Perninder Dhadwar, Imobisoft
- Mark Wilson, Headpoint Advisors
- Rachael Essex, Headpoint Advisors
- David Bellis, BGF
- Professor Abdul Sadka, Aston University
- Alexandra Kalu, Fieldfisher
- Yadu Szmulewicz, Conigital
- Richard Smith, PhysioGenie Technology
The sections below summarise the AI challenges in business the panel discussed, as we reported them in June 2024. Insider Media also published its own write-up of the round table.
Why does AI adoption need a clear problem to solve?
David Bellis, an investor with BGF, pointed out that for AI-powered businesses to attract investment, they must clearly identify and address specific problems that AI can solve, ensuring a ready customer base for their solutions. Perninder Dhadwar emphasised the importance of problem-solving over technology for its own sake.
In fact, the government data backs them up. Not having identified a use for AI was the top barrier in DSIT's survey, ahead of skills, tools and cost. A business that cannot name the process it wants to improve cannot judge whether an AI project worked. That is why we put AI strategy before code.
Why is AI adoption harder for SMEs?
Perninder Dhadwar, CEO of Imobisoft, noted that while larger companies often push for AI adoption, smaller businesses face challenges due to high investment costs and the current limitations of AI.
The ONS figures show the gap. Around 28% of businesses with fewer than 10 employees use AI, against 49% of those with 250 or more. Only around 2% of the smallest businesses are recruiting AI-skilled staff, compared with 10% of those with 250 or more.
Cost plays its part. For example, one small business told DSIT's researchers: "The challenge is mostly cost. You don't know if you're buying something fantastic or not." For a small firm, a failed pilot takes a bigger share of the budget. Therefore, it makes sense to test AI on one narrow process before committing more. Our guide to AI on a smaller budget covers lower-cost ways to start.
What is the AI skills gap?
The AI skills gap is the difference between the AI skills a business needs and the skills its people have. Professor Abdul Sadka, Director of the Aston Digital Futures Institute, highlighted the importance of ongoing skills development in AI. As AI technologies advance rapidly, there is a pressing need for continuous upskilling and reskilling to keep the workforce adept and relevant. Universities, he said, play a key role in providing this training through degree apprenticeships and continuous professional development programmes. His institute has since been renamed the Sir Peter Rigby Digital Futures Institute.
Similarly, Alexandra Kalu from Fieldfisher said that in the legal sector, automating routine tasks raises productivity. However, the shift requires significant upskilling and training of staff.
The gap has not closed since the round table. Limited AI skills were the second most common barrier in DSIT's survey, cited by 60% of businesses. The ONS found only 11% of firms with 10 or more employees have trained over half their staff in AI. Training existing staff is the most common response, reported by around 40% of medium and large businesses.
Can businesses rely too much on AI?
Mark Wilson, Managing Partner at Headpoint Advisors, cautioned against over-reliance on AI, especially for customer-facing decisions, highlighting the need for experienced professionals to interpret AI outputs and ensure quality control.
Two other speakers also raised a longer-term risk. Yadu Szmulewicz of Conigital warned that over-reliance on AI by junior staff could lead to a future lack of experienced professionals capable of critical evaluation. Perninder Dhadwar highlighted the risk of losing expertise if junior staff are not adequately trained because AI has taken over basic tasks.
The concern is also practical. In DSIT's interviews, the accuracy of AI output was one of the most common challenges in deploying AI. It came up particularly among smaller and micro businesses. Someone in the business has to be able to tell when the output is wrong.
The trap
Handing routine work to AI and removing the tasks through which junior staff learn the job.
The fix
Let AI draft the routine work, have junior staff check and correct it, and keep experienced people signing it off.
The same applies in software development, where AI-generated code still needs verification before it reaches production.
What are the data privacy risks of using AI providers?
Alexandra Kalu raised concerns about data privacy and ethical considerations when using third-party AI providers. She stressed the importance of maintaining client trust. When client data passes to an outside AI service, the firm also relies on that provider's terms and security.
Likewise, DSIT's research reflects the same worries. Data security was one of the most common challenges in deploying AI safely, especially for larger firms. In addition, 22% of businesses cited ethical concerns as a barrier to adoption. For sensitive data, a privately hosted model can be the better choice. Our guide explains when a private LLM beats an API.
Building trust and keeping a human element
Rachael Essex of Headpoint Advisors stressed the importance of building trust in AI technologies among the general public. Despite AI's capabilities, she said a human touch still matters in client interactions and marketing. It helps ease fears and scepticism about AI and big data.
Her point concerns the people on the receiving end of AI. However, most businesses in DSIT's interviews mentioned no case of trust, reliability or safety concerns affecting their AI decisions. Clients and the public, on the other hand, are a different audience. Our framework for responsible AI covers how to explain AI decisions to the people they affect.
What opportunities did the panel see in AI?
Mark Wilson emphasised that AI's primary value lies in automating repetitive, time-consuming tasks, freeing people for more strategic and complex problem-solving. He added that this shift could also change billing practices, from time-based to solution-based payments. Not every repetitive task needs AI, though. Rule-based tools are often cheaper, as our comparison of AI vs automation explains.
Alexandra Kalu described how AI is improving operational efficiency within the legal sector.
Yadu Szmulewicz and Richard Smith of PhysioGenie Technology underlined the need to demonstrate real-world benefits and efficiencies from AI. Smith added that new data collection methods, such as gamification, can improve AI training and development.
Despite talk of AI fatigue, David Bellis believed we were only beginning to see AI's potential. Many businesses, he said, were integrating AI to enhance their existing operations. Indeed, the numbers since then support him. The ONS puts AI use among firms with 10 or more staff at around 35% in June 2026. That is up from around 12% in late 2023.
How to tackle these AI adoption challenges
The round table showed how widely AI and digital transformation are reshaping business. While AI offers significant opportunities for improving efficiency and solving complex problems, it also presents challenges related to skills development, ethical considerations and maintaining trust.
The data since then suggests an order for tackling them:
- Name the problem first. Not having found a use for AI was the top barrier in DSIT's survey.
- Budget for training alongside the software, since skills were the second biggest barrier.
- Keep experienced people checking AI output, especially where customers are affected.
- Check where your data goes before using a third-party AI service.
These are the challenges of AI in business that decide whether a project pays off. Once AI is in place, the next question is the return. We cover that in our look at where AI's business value will come from next.
Not sure where AI fits in your business?
We help UK businesses find the process where AI will pay off. We will also tell you when simpler automation would do the job.
Our free 30-minute discovery call looks at one process in your business and gives you a rough cost, timeline and estimated return.
Book a discovery callFrequently asked questions
What are the biggest AI adoption challenges?
In DSIT's survey, the top barrier was not having identified a use for AI, cited by 71% of businesses. Limited AI skills (60%) and a lack of tools (48%) came next, followed by unclear regulation, cost and ethical concerns.
Why is AI adoption harder for small businesses?
Smaller firms have less to spend and fewer specialist staff. In the ONS's June 2026 survey, 28% of businesses with fewer than 10 employees used AI. The figure for businesses with 250 or more was 49%.
What is the AI skills gap?
It is the gap between the AI skills a business needs and the skills its staff have. Limited AI skills were the second most common barrier to AI adoption in DSIT's survey, cited by 60% of businesses.
What are the risks of over-reliance on AI?
Mistakes can go unchecked, particularly in customer-facing decisions. Junior staff may also miss out on the experience needed to judge AI output. Keeping experienced people reviewing the results reduces both risks.
How do digital transformation and AI fit together?
Digital transformation means changing how a business works using technology. The role of AI in digital transformation is to handle tasks that fixed rules cannot, such as predictions or unstructured information.
Sources
- AI adoption survey report 2025, Department for Science, Innovation and Technology, 28 January 2026
- Artificial intelligence in UK businesses: 2023 to 2026, Office for National Statistics, 20 July 2026
- Growth through Digital & AI, Insider Media
- The Sir Peter Rigby Digital Futures Institute, Aston University