13 employees built 53 AI tools and recovered 3,783 hours

Thirteen people. Fifty-three working AI tools. Nearly 3,800 hours of recovered annual capacity. That is the headline from one US property company's internal AI programme, and the number that should make UK operations managers, HR heads, and managing directors sit up is not the tool count. It is the fact that none of those thirteen people were developers. They were ordinary team members who learned to build AI tools themselves, without writing a single line of code. The real story is what that says about how AI adoption actually works.

The problem most teams recognise immediately

Teams were spending too much time on work that did not need their expertise. Repetitive data entry, basic analysis, reformatting documents, chasing information across systems: the kind of work that fills the day and empties the diary of anything more valuable. This is not a technology problem. It is a workflow problem that technology can address, but only if the people doing the work understand the tools well enough to apply them.

For many businesses, the default response has been to buy a platform and hope adoption follows. It rarely does. Staff use 20% of the features, revert to old habits within weeks, and the investment produces a report rather than a result. Operations managers and managing directors will recognise this pattern immediately.

What the AI programme actually did

Rather than rolling out a single enterprise platform, the business ran a structured internal programme in which thirteen employees built AI tools to address their own workflow challenges. The programme was scoped around practical application, not theory. Participants identified where repetitive work was costing time, designed a tool to address it, and built it themselves using no-code or low-code methods.

The result was 53 completed tools covering tasks across multiple departments. These were not prototypes or proofs of concept filed away in a shared drive. They were working tools embedded into daily operations.

Why the results matter for UK SMEs

The headline numbers are significant. 3,783 hours of recovered annual capacity is equivalent to roughly 473 working days. That is roughly two full-time roles redirected from routine work to higher-value activity. Self-assessed output quality rose from 3.2 to 4.3 on a five-point scale, a 34% improvement in how staff rated the work they were producing.

But the more important number is thirteen. That is how many people drove those results. For a UK SME with a team of 20 to 100, this is not an enterprise case study viewed from a distance. It is a directly comparable model. A small cohort of trained, motivated team members can build AI capability that pays back across the whole business. The same model fits a fire protection firm, a trade association or a law firm, where the people closest to the paperwork know best where time goes.

Here is where gecco's defining view applies: AI adoption is 80% people and culture, 20% technology. The tools in this programme did not succeed because they were sophisticated. They succeeded because the people building them understood the problem they were solving, had the confidence to build, and were given the structure to do it. Managing directors who focus on the tool count are measuring the wrong thing. The measure that matters is staff confidence and the quality of the work those tools replace.

What this approach looks like in practice

For an operations manager or HR lead thinking about how to replicate this, the programme structure is more instructive than the output numbers. Three elements made it work.

First, the scope was narrow and practical. Participants were not asked to think about AI in the abstract. They were asked to identify one or two tasks in their own role that cost time without adding much value, then build something to address those tasks specifically.

Second, the skills transferred before the tools were built. Staff were given enough grounding in how AI tools work, what they can and cannot do, and how to prompt them effectively before they started building. That sequence matters. Confidence in the tool comes from understanding it, not from using it.

Third, the output was measured. Output quality was tracked using a consistent self-assessment scale, giving the programme a way to demonstrate value that went beyond the number of tools completed. That is important for managing directors who need to justify AI investment to a board or to themselves.

Implementation safeguards worth noting

Any UK business running a similar programme needs to address three practical constraints before participants start building.

First, define acceptable use clearly. Staff building their own tools will make decisions about what data goes into a prompt. Without a written policy, those decisions will be inconsistent and potentially expose the business to risk.

Second, apply GDPR controls if staff or customer data is used in prompts or training. This is not optional. The ICO has published guidance on AI and personal data, and UK SMEs are within scope.

Third, build in output review at the start. AI-generated content and analysis should be checked by a human before it is used in a decision, a client document, or an external communication. Self-assessed quality improvements are a useful internal measure, but they do not replace editorial judgement.

These safeguards are not reasons to delay. They are the conditions that make a programme sustainable rather than one that runs well for three months and then quietly collapses when something goes wrong.

Making this work for your business

The model described here is accessible to UK SMEs. It does not require a large technology budget or a dedicated AI team. It requires a structured training approach, a clear scope, and the willingness to give a small cohort of staff the time and permission to build.

The practical starting point is identifying which roles in the business carry the most repetitive, lower-value work. Operations and finance functions are usually the highest-return areas. HR processes, from onboarding documentation to policy queries, are also strong candidates. The key question for an operations manager or HR head is not which AI tools to buy. It is which tasks, if automated or assisted, would free the most capacity for work that actually needs human attention.

gecco's Training and consultancy programme gives SME teams the skills to identify, build, and embed AI tools into their own workflows, with a 92% onboarding cohort completion rate that reflects how the programme is structured around practical application from day one.

Start by understanding where your business stands

If you are an operations manager, HR head, or managing director weighing up whether an internal AI programme could work for your team, the AI Readiness survey is designed to surface exactly that question.

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 30-minute showcase to walk through the results with a gecco adviser.

---

Website · LinkedIn · Case Studies · Newsletter