How UK SMEs are closing the AI operations gap

AI adoption among UK businesses has nearly tripled in under three years. That headline figure is striking. What sits beneath it is more instructive: the businesses leading this shift are not deploying experimental technology. They are using AI to fix the same operational bottlenecks that operations managers and business owners have complained about for years. Manual data entry, slow reporting, fragmented systems, decisions made on yesterday's numbers. The gap between early adopters and everyone else is widening. Understanding what the adopters are actually doing, and why it works, is the more useful question.

The scale of the shift in UK AI adoption

Official ONS data, tracking AI use in UK businesses from 2023 to 2026, show that self-reported AI use among firms with 10 or more employees rose from around 12% to around 35% between late 2023 and June 2026. That is not gradual drift. That is a competitive field reshaping itself in real time.

Large language models are the most widely used AI technology, at 18% of businesses with 10 or more employees as of June 2026. These are not specialised enterprise deployments. They are mainstream tools that any team member can access today.

The most reported use of AI across all adopting businesses is improving operations. Close to 60% of businesses using AI cite this as a primary application. For operations managers evaluating where to start, that consensus is a useful signal.

Where smaller firms are falling behind

The ONS data also reveal an adoption gap. Larger firms are more likely to adopt AI, and smaller firms remain underrepresented in the adoption figures. Many businesses in the 10 to 50 employee range continue to rely on manual processes for tasks that AI handles reliably elsewhere.

This matters for business owners running lean teams. When a competitor with similar headcount automates its reporting, speeds up its data consolidation, or uses AI to personalise its customer offer, the gap is not just operational. It shows up in response times, cost structure, and decision quality.

Around 31% of the smallest firms report using AI to personalise products or services. That figure suggests a meaningful share of small businesses have moved beyond experimentation. They are applying AI where it directly affects revenue and customer experience.

Why this is 80% a people problem, not a technology problem

At gecco, our view is that AI adoption is 80% people and culture, and 20% technology. The ONS data support this framing. The most common adoption route among UK businesses is training existing staff rather than hiring new AI specialists.

This is the pragmatic choice for most SMEs. Hiring a dedicated AI specialist is expensive and slow. Training the operations manager who already understands the business processes is faster and often more effective. But it only works if the upskilling is structured.

Shallow upskilling, a one-day workshop, a few tool demonstrations, a loose encouragement to experiment, produces shallow results. Staff return to existing habits. AI use remains patchy. The operational gains that justify the investment never materialise consistently.

The businesses achieving sustained efficiency gains are those that give their teams a clear framework for where AI fits, what it is expected to do, and how to evaluate whether it is working. That is not a technology decision. It is a management and training decision.

What structured AI adoption looks like in practice

For operations managers and business owners, the practical question is not whether to use AI. It is which processes to automate first, and how to build the internal competence to sustain those automations.

The most productive starting points tend to share three characteristics. First, they involve high-volume, repetitive tasks where the inputs and outputs are clearly defined. Data consolidation from multiple systems, routine reporting, and status updates are typical examples. Second, they connect existing tools rather than replacing them. Linking a CRM to an accounting platform to remove a manual re-entry step is lower risk than deploying a new system. Third, they are visible enough that team members can see the time saved. Early wins build the internal appetite for broader adoption.

Businesses that approach AI adoption this way tend to generate compounding returns. Each automation frees up time. That time gets reinvested in the next improvement. The team builds confidence and competence together, rather than one specialist carrying the entire AI workload.

Considerations and realistic limits

The ONS figures measure self-reported use, not verified outcomes. Some of the businesses counted in the 35% figure may be using AI tools lightly, without the structured approach needed to generate consistent operational gains. Adoption rate and adoption depth are different measures.

For operations managers scaling AI use across a team, UK data protection law applies to any process that handles personal data. The ICO publishes guidance on AI and data protection that is worth reviewing before automating customer-facing processes. Emerging government guidance on responsible AI deployment is also evolving, and businesses scaling AI should monitor it.

There is also an honest point about staff readiness. Not every team member will adopt new tools at the same pace. Rushing automation into a team that has not been prepared for it creates resistance, errors, and, eventually, abandoned workflows. The sequence matters: train first, automate second, embed third.

Making this work for your business

For operations managers and business owners, the data point to a clear and practical starting position. Identify the two or three manual processes that consume the most time across your team. Assess whether the inputs and outputs are consistent enough to automate. Then build the internal competence to run and maintain those automations before expanding further.

Structured training is what separates the businesses achieving lasting gains from those stuck in pilot mode. gecco's Training and consultancy programme gives operations managers and business owners a vendor-neutral framework for exactly this: understanding where AI fits in your specific workflows, building assistants that reflect your business context, and developing the internal competence to sustain adoption without ongoing dependency on external support.

Ready to find out where AI fits in your operations

If you are an operations manager or business owner weighing up which processes to automate first, and whether your team has the foundations to make it stick, the AI Readiness survey is built to surface exactly that.

Take the AI Readiness survey. You will receive 65+ free resources and a custom AI Readiness report based on your answers. From there you can book a free 45-minute AI Readiness call to walk through the results with a gecco advisor.

Website · LinkedIn · Case Studies · Newsletter