How a World-Class Isle of Man Manufacturer Turned to Machine Learning to Transform Its Production Line
Back to all posts
Case Study15 March 202615 min read

How a World-Class Isle of Man Manufacturer Turned to Machine Learning to Transform Its Production Line

Strix Plc, Insight Innovation, and the Isle of Man Government’s Activate AI Programme come together to deliver a pioneering AI project — and name it after a beloved former colleague.

When Strix Plc’s Ramsey facility celebrated the completion of its first artificial intelligence project, the occasion was marked by an extraordinary gesture. Barbara Hay, a former member of staff who had retired from the company twenty years earlier, was invited back to the factory floor. The machine learning tool that now helps guide the production line had been named in her honour: AI-Barbara.

It was a moment that captured something important about this project — and about the Isle of Man’s approach to AI more broadly. This was not a story about technology replacing people. It was a story about people building technology that honours the knowledge and experience accumulated over decades, and using it to make a world-class manufacturing operation even better.

Strix: A Global Leader Rooted in the Isle of Man

Strix Group Plc is one of those rare companies that most people rely on every day without ever knowing its name. Founded on the Isle of Man in 1982 by Dr John Taylor, Strix is the global leader in the design, manufacture, and supply of safety controls for kettles and other small domestic appliances — precision-engineered components trusted by brands including Philips, Tefal, Russell Hobbs, Morphy Richards, Braun, and Bodum. The company estimates its controls are used approximately 1.2 billion times per day, in over 100 countries, by more than ten per cent of the world’s population.

The company’s headquarters remain at Forrest House, Ronaldsway, on the Isle of Man, with manufacturing operations in Ramsey, Guangzhou in China, Vicenza in Italy, and Melbourne in Australia. Listed on AIM (London Stock Exchange: KETL), Strix employs over 900 people globally and has manufactured more than two billion safety controls since its founding.

The Ramsey facility is home to one of the company’s most demanding production environments: high-speed, tight-tolerance manufacturing lines producing precision components at scale. The nature of this work demands extraordinary accuracy. The tooling set-up for each production run must be calibrated to account for the specific characteristics of each incoming batch of raw material. Even small variations in the material can require adjustments to the machine settings to ensure that finished components meet Strix’s exacting quality standards.

Historically, this calibration process has relied heavily on the skill and experience of the production team — the setters who operate the lines. It is painstaking, expert work. Getting the machine settings right for a new batch of material could take up to four hours of trial and adjustment. During that set-up period, the line is not producing saleable product, and the material used in the calibration process is typically scrapped. In a high-volume manufacturing environment producing millions of components annually, those hours of downtime and that scrap material add up to a significant cost.

This was the challenge that Insight Innovation was brought in to address.

Insight Innovation: Applied Physics Meets Machine Learning

Insight Innovation is an AI and automation consultancy based on the Isle of Man. Founded by Colin Moughton — an applied physics modeller and inventor listed on over one hundred patents — the company occupies a distinctive position in the island’s growing AI ecosystem. Colin has spent his career computationally modelling complex physical systems, and his PhD involved modelling the very production processes used at Strix Ramsey. It is a background that gave him an unusually deep understanding of both the science underlying the manufacturing operation and the data it generates.

While many AI consultancies focus on training, workshops, and the deployment of off-the-shelf tools, Insight Innovation builds bespoke machine learning and custom software solutions from the ground up. The company’s approach is rooted in deep domain expertise: understanding the physics, engineering, and operational realities of the problems it is asked to solve, and then designing and building the AI systems that address them.

As an official Activate AI partner through Digital Isle of Man, Insight Innovation is part of the Isle of Man Government’s strategic initiative to accelerate AI adoption across the island’s economy. But the Strix project stands out because it goes beyond awareness-raising and tool deployment. It represents the kind of applied, production-grade machine learning that delivers measurable operational impact in a demanding industrial setting.

It is the combination of domain knowledge and AI capability that makes this project unusual, and that made its success possible.

The Project: From Data to Decisions

The project was delivered in two phases.

Phase One: Proof of Concept

The first phase set out to establish whether a machine learning model could usefully predict the optimal tooling set-up for the tight-tolerance production lines. Working with historical production data — summarised records of previous batch runs, material properties, and machine settings — Colin developed a neural network using TensorFlow, one of the leading open-source machine learning frameworks. The model was trained to learn the relationships between incoming material characteristics and the machine settings that had historically produced the best results.

The user interface was built using Streamlit, a Python-based framework for creating interactive data applications. This gave the production team — the setters and engineers on the factory floor — a straightforward, visual way to interact with the model: input the characteristics of the new material batch, and receive a recommended machine setting.

The first phase demonstrated the viability of the approach. The model could identify meaningful patterns in the historical data and produce useful recommendations. However, it also revealed an important limitation. The summarised historical data on which the model was trained lacked the granularity needed to make highly accurate predictions across the full range of material variations the line encounters. The data captured the broad strokes of previous production runs, but not the fine-grained detail that would allow the model to distinguish between subtle material differences with confidence.

This is a common and critically important finding in applied machine learning. The model is only as good as the data it learns from. The first phase had proven the concept; the second phase would need better data.

Phase Two: Real-Time Data and Refined Predictions

For the second phase, Colin proposed a fundamental improvement to the data pipeline. Rather than relying on summarised historical records, the project needed access to real-time production data — granular, time-stamped measurements from the production line itself, captured as the machines operated.

Working closely with the Strix team — Anne from the production department and Ryan from IT — Colin designed and implemented a live data feed that connected the production line’s monitoring systems directly to the TensorFlow predictive model. This was careful, detailed work. It required understanding the data architecture of the production systems, establishing secure and reliable connections, handling data formats, managing the flow of information in real time, and ensuring that the solution integrated cleanly with the factory’s existing IT infrastructure.

The result was transformative. With access to far more granular, real-time data, the model’s predictions improved dramatically. AI-Barbara could now analyse the specific characteristics of each incoming material batch with much greater precision and recommend machine settings that were, in many cases, right first time.

The impact was immediate and measurable. Set-up times that had previously taken up to four hours were reduced to right-first-time results. The reduction in scrap material and downtime is projected to yield an additional 4.3 million components per year — approximately £50,000 in annual savings. In a precision manufacturing environment, these numbers represent a meaningful operational improvement: more product from the same line, less waste, and production teams freed to focus their expertise on the work that most benefits from human skill and judgment.

The Technology Behind AI-Barbara

For those interested in the technical architecture, the solution is built on a straightforward but powerful stack.

At its core is a TensorFlow neural network — a type of machine learning model well suited to the kind of multivariate regression problem that tooling calibration represents. The model takes as input a set of material properties and production parameters, and outputs predicted optimal machine settings. It was trained on labelled historical data in Phase One and then re-trained on the far richer real-time data stream in Phase Two, with the improved dataset producing markedly better prediction accuracy.

The entire application is written in Python, leveraging the scientific computing ecosystem — NumPy, Pandas, and TensorFlow — that has made Python the dominant language for applied machine learning. The user-facing interface is built in Streamlit, which allows rapid development of interactive web-based tools without the overhead of a traditional web application framework. This was a deliberate design choice: the interface needed to be intuitive enough for production setters to use confidently on the factory floor, without requiring any specialist software knowledge.

The data pipeline developed in Phase Two connects the production line’s monitoring systems to the model in near real-time, ensuring that predictions are always based on the most current and relevant data available.

Government Support: The Department for Enterprise Business Consultancy Scheme

The second phase of the project was part-funded through the Isle of Man Department for Enterprise’s Business Consultancy Scheme. This scheme offers Isle of Man–registered businesses a grant of up to fifty per cent of the cost of external consultancy projects, to a maximum of £7,500 per phase, in areas that create jobs, increase revenues, or improve efficiency.

The Business Consultancy Scheme is one of a suite of enterprise support programmes administered by the Department for Enterprise, which has a mission to be a forward-thinking partner supporting island businesses and people to fulfil their potential. The Department oversees four executive agencies — Digital Isle of Man, Finance Isle of Man, Business Isle of Man, and Visit Isle of Man — each working to promote sustainable economic growth across the island’s key sectors.

The Strix project represents an important milestone for the scheme in the context of AI. It demonstrated that the Business Consultancy Scheme can be used effectively to support artificial intelligence and machine learning projects — not just traditional business consultancy disciplines. Strix and Insight Innovation worked collaboratively with the Department to navigate the application process, and the experience provided valuable learning for all parties on how the scheme can accommodate the distinctive characteristics of AI project delivery.

Nick Gibbs, Engineering Director at Strix, championed the application from the company side, recognising both the operational value of the project and the opportunity to demonstrate that Isle of Man manufacturers can access government support to adopt cutting-edge technology. The grant was approved for the second phase, reflecting the fact that this was a continuation of proven work with clear, measurable objectives.

The success of this funding pathway is significant. It establishes a precedent for other Isle of Man businesses considering AI adoption. The message is clear: government support is available, the process works, and the results can be substantial.

The National AI Office and the Isle of Man’s AI Ambitions

The Strix project did not happen in isolation. It is part of a much larger story about the Isle of Man’s determination to position itself at the forefront of practical, responsible AI adoption.

In January 2026, the Isle of Man Government launched its National AI Office (NAIO), backed by £1 million in public funding. Led by Digital Isle of Man — an executive agency of the Department for Enterprise — the NAIO brings together existing functions and expertise from across government and industry to accelerate responsible AI adoption, strengthen governance, and support innovation that benefits the island’s economy and public services.

The NAIO builds on the success of the Activate AI programme, launched by Digital Isle of Man in August 2024. In its first year, the programme engaged nearly 1,000 participants in AI training sessions and generated an estimated £2 million in productivity savings across the island’s economy. It established a network of Activation Partners — vetted local AI specialists, including Insight Innovation — who work directly with businesses to identify opportunities, develop solutions, and trial AI-driven tools.

The programme’s ambition is striking: to increase the Isle of Man’s GDP by ten per cent by 2030 through AI-driven solutions across both the public and private sectors. The NAIO’s first-year priorities include finalising a National AI Strategy, delivering an island-wide AI literacy programme, producing guidance on safe and responsible AI use, applying AI to public service reform, and assessing workforce impacts including reskilling and training needs.

Enterprise Minister Tim Johnston has described the initiative as a natural next step, building on the foundations already laid by Digital Isle of Man and its partners. The emphasis throughout is on practical adoption — helping real businesses solve real problems — rather than abstract strategy. The Strix project embodies precisely this philosophy.

For international observers, the Isle of Man’s approach is noteworthy. Here is a small, agile jurisdiction — a self-governing Crown Dependency in the Irish Sea — that has committed significant public resources to AI, built a structured programme connecting businesses with specialist partners, created funding pathways to reduce the cost of adoption, and is now seeing tangible results in world-class manufacturing operations. It is a model that larger jurisdictions could learn from.

What This Means for Isle of Man Businesses

The Strix and Insight Innovation project offers several important lessons for other businesses on the island — and beyond.

AI works in manufacturing. This is not a technology reserved for Silicon Valley software companies. A precision manufacturing facility on the Isle of Man has deployed a machine learning model that delivers measurable, daily operational improvements. The tools and frameworks used — Python, TensorFlow, Streamlit — are mature, well-supported, and accessible.

Domain expertise matters as much as AI expertise. The project succeeded in large part because the person building the model had spent years computationally modelling the physics of the production process. AI applied without domain knowledge often produces models that look impressive on paper but fail in practice. The combination of deep manufacturing knowledge and machine learning capability was essential.

Data quality is everything. The difference between Phase One and Phase Two of this project was not a more sophisticated algorithm. It was better data. The move from summarised historical records to a real-time data pipeline transformed the model’s accuracy. Any business considering AI should start by understanding what data it has, how good that data is, and what it would take to improve it.

Government support is available and it works. The Department for Enterprise’s Business Consultancy Scheme can fund AI projects. Strix has demonstrated the pathway. Businesses interested in exploring AI-driven improvements can apply for match funding that significantly reduces the cost of engagement with specialist consultancies like Insight Innovation.

AI augments people — it does not replace them. AI-Barbara does not replace the setters who operate the tight-tolerance lines. It gives them a powerful tool that draws on decades of accumulated production data to recommend a starting point. The setters bring their skill, judgment, and experience to the process. The result is a team that is faster, more efficient, and producing less waste — not a smaller team.

Looking Ahead

The success of AI-Barbara opens the door to further applications of machine learning at Strix and across the Isle of Man’s manufacturing sector. The data infrastructure built during Phase Two — the real-time pipeline connecting production systems to predictive models — is a foundation that can be extended. Predictive maintenance, quality anomaly detection, yield optimisation, and supply chain forecasting are all areas where similar approaches could deliver significant value.

For Insight Innovation, the project is a proof point for the kind of work the company was built to do: solving real engineering and manufacturing challenges with bespoke AI systems, not generic tools. As the Isle of Man’s AI ecosystem matures — supported by the NAIO, the Activate AI programme, and the Department for Enterprise’s funding schemes — the demand for this level of specialist capability is likely to grow.

For the Isle of Man Government, the Strix project is exactly the kind of success story that the Activate AI programme and the NAIO were designed to catalyse. A globally significant manufacturer, headquartered on the island, working with a local AI specialist, supported by government funding, and delivering quantifiable improvements to a high-technology production process. It is a powerful demonstration of what practical, well-supported AI adoption looks like — and an invitation to other Isle of Man businesses to explore what AI could do for them.

And for Barbara Hay, invited back to the factory floor two decades after her retirement to meet her digital namesake, it was a reminder that the knowledge and experience of the people who built Strix’s reputation for excellence live on — now encoded in a machine learning model that helps the next generation of setters continue that tradition.

---

Strix Group Plc (AIM: KETL) is the global leader in the design, manufacture, and supply of kettle safety controls, with headquarters on the Isle of Man and operations worldwide.

Insight Innovation is an AI and automation consultancy based on the Isle of Man, specialising in machine learning, custom software, and data-driven solutions for engineering and manufacturing businesses. An official Activate AI partner through Digital Isle of Man.

The Department for Enterprise’s Business Consultancy Scheme offers Isle of Man–registered businesses grant funding of up to 50% towards the cost of external consultancy projects.

The National AI Office (NAIO), led by Digital Isle of Man, coordinates AI policy, adoption, and governance across the Isle of Man. For more information about the Activate AI programme and NAIO, visit digitalisleofman.com.

Want to discuss how AI could work for your business?

Book a free 30-minute discovery call. No pitch. No pressure.

Book a Discovery Call