AI Adoption in UK SMEs: Why 2025 Adoption Was Slow And 2026 Is Already Different

AI is transforming almost every sector of the economy, but for UK SMEs, the reality in 2025 was far more modest than the headlines suggested. In February this year, the Department for Science, Innovation & Technology (DSIT) published a government review of AI adoption to date, built on 3,500 business interviews and 100 qualitative follow‑ups conducted between February and May 2025.*

The results were striking, not because adoption was high, but because it was surprisingly low. However, this data already feels out of date. From what I see in early 2026, the acceleration of AI in businesses is already underway.

In this blog, I reflect on what the government’s 2025 data really tells us about SMEs, why I believe 2026 is shaping up very differently, and how we at TC Group are building the foundations to support both our people and our clients through this shift.

2025: A YEAR OF LOW AI ADOPTION ACROSS UK SMES?

The government’s 2025 research paints a clear picture: AI adoption remains modest. The data showed that only 1 in 6 businesses currently use AI, and most had no active plans to adopt it in 2025.*

The qualitative and quantitative insights showed that:

• Most SMEs had not deployed AI systematically or strategically.
• Adoption was patchy, small‑scale, and usually confined to experiments.
• Many business leaders remained uncertain about where AI could be applied.

The research highlighted a crucial underlying issue: most SMEs struggled to identify clear AI use cases, and the study carefully noted that it did not capture “shadow AI” (the unofficial, unmanaged use of consumer AI tools by employees).

In other words, even where AI was being used, it often wasn’t part of formal business operations.

As someone who works with SMEs daily, I was not surprised by any of this. In 2025, most SMEs were still caught between hype and hesitation.

 

WHY WAS 2025 AI ADOPTION IN SMES SO LOW?

From my conversations with SME owners, the barriers identified by the government resonate strongly:

1. Lack of clarity about what AI could help with

The government findings stress the need for “more detail” on real business use cases, because most SMEs don’t know how AI connects to their workflows or data structures.

2. Difficulty accessing or joining up data

Even small businesses now sit on messy, siloed or legacy datasets. Without coherent, connected data environments, AI simply cannot deliver scalable results.

3. Limited internal capability

The 2025 research emphasised the importance of digital infrastructure and skills – areas where many SMEs felt underprepared and under‑resourced.

When you combine these factors, “modest” adoption is exactly what you’d expect.

 

2026: A YEAR OF WIDER AI ADOPTION?

Despite the sluggish picture in 2025, I’m already seeing an acceleration in early 2026, driven by three major shifts:

1. SME leaders are beginning to understand AI better

Even in the past 12 months, awareness has deepened significantly. Tools like Microsoft Copilot, ChatGPT, and sector-specific AI assistants have become mainstream, helping to demystify AI for business owners.

2. Vendors are maturing, and business solutions are becoming clearer

In 2025, vendors mostly sold demos. In 2026, they’re beginning to understand that they need to sell use‑case‑led solutions.

This aligns fully with what we’ve been working on in TC Group. We understand that we cannot build a business strategy around AI demos or individual products; we can only build one around AI that solves real, well-defined business objectives.

3. Data foundations are finally being laid for AI to work

I’m seeing SMEs begin to lay better foundations for AI to work. At TC Group, 2025 was the year we did the heavy lifting:

• Integrating core systems and datasets across practice systems and CRM.
• Centralising workflows.
• Building unified datasets.
• Establishing a central digital environment where AI can operate.
• Creating TC Digital LABS, our controlled, secure environment for experimenting with AI tools, testing new ideas and evaluating emerging technologies before they go live.

This groundwork allows us to now roll out AI initiatives at pace. I’m starting to see the same thing in the SMEs we support. Once the data structure stabilises, AI stops being theoretical and becomes operational.

2026 is the year AI gathers operational pace

Once the objectives are defined and the layers of data foundations are in place, SMEs will be able to run AI development in sprints.

 

AI ALONGSIDE ENHANCED HUMAN CAPABILITY

The government’s research emphasised the UK’s goal to embed AI across the economy to drive productivity and competitiveness. SMEs, however, won’t get there through automation alone. At TC Group, we understand that AI will only improve our client experience when combined with enhanced human capabilities.

In many sectors, AI may provide speed and scale, but humans still provide expertise, empathy and real interactions.

In our industry, the role of the accountant is shifting. We’re seeing AI as a force multiplier for human expertise. We’re already seeing this in TC Live Accounting, where automation is freeing teams to spend more time understanding clients’ aspirations, constraints, and growth opportunities, and advising them.

 

SUMMARY

The government’s 2025 research made clear that AI adoption in UK SMEs was low, modest, fragmented, and often superficial.

But based on what I see happening now, both in our firm and across our client base, 2026 will not follow the same trajectory.

I see 2026 being fundamental to SMEs where:

• AI knowledge is increasing.
• Technology is stabilising.
• Data foundations are being created.
• Vendors are more focused on business objectives and solutions.
• SMEs are moving from curiosity to practical adoption.

At TC Group, we’re committed to supporting that journey, not only through our own internal AI evolution but also ultimately empowering our clients to understand, adopt, and benefit from AI safely and meaningfully.

 

*DATA SOURCE: UK GOVERNMENT AI ADOPTION RESEARCH Published 13 February 2026