Engineering Intelligence: AI at IMI

"We think about AI through three different lenses," says Luke. "The first is how we drive value for our customers and provide excellent service to them. The second is growth and how we evolve business models around AI for the longer term. The third is operational and personal productivity. "

This approach reflects our measured approach to recent developments in AI technology that will ultimately be transformative for every business but that require discipline and controls to be deployed effectively. As a global leader in fluid and motion control, we operate in environments that are safety-critical, producing components where there is zero margin for error. This means we must balance our pace of change with a focus on rigour and governance.

Prioritising investment

Having a clear process for prioritising where and how AI can improve workflows and customer outcomes is crucial. We employ a risk-based decision matrix that maps two factors: the cost of error and the level of tacit knowledge required for a task. Tacit knowledge is intuitive and experience-based, shaped by personal judgement and context. Explicit knowledge is rule-based and factual, meaning it can be more easily documented or programmed. Explicit knowledge is therefore easier to replicate through automation.

"If a task involves very little tacit knowledge and the cost of error is very low, it is a strong candidate for automation," Michael explains. "At the other end of the spectrum, where a task requires significant tacit knowledge and the cost of error is high, for example because it is safety-critical, then it needs to remain human-led."

Between these two extremes lies the territory where AI creates value through augmentation rather than automation. "For these processes, AI helps us work more efficiently, but there is always a human validating and making the final decision," Michael says.

Balancing governance and innovation

With operations spanning multiple sectors and geographies, we face a common challenge: how to fuel AI innovation across the group. The solution is to focus on parallel centralised and decentralised capabilities.

In practice, that means that central teams develop major applications designed to work across multiple sectors and business functions to ensure security, governance and cost control. Meanwhile, individual sites and teams are encouraged to experiment with AI tools for their own productivity and local workflows. Luke emphasises why this works, "Our culture is to encourage entrepreneurship at a local level," he says.

IMI uses a blend of in-house and external expertise to get the benefit of new ideas and the control and security of in-house capability. Michael explains that this is best practice: “We have a strong internal Data & AI team and a secure environment for building AI products. Where helpful, we also partner with third parties to bring new perspectives and specialist expertise, working closely with our team and within our secure infrastructure.”

Governance is crucial. It encompasses not only strategic alignment, but also the safe, ethical and responsible use of AI, alongside cybersecurity, privacy and regulatory compliance. At IMI, our AI steering committee, comprising IMI’s Executive Team plus specialists from around the business, reviews major investment decisions and ensures that our priorities are being delivered through AI innovation.

Applications in practice

To drive commercial excellence, we are using AI to improve our customer service through faster response times and more efficient processes. "Many IMI products have data sheets. Being able to analyse these quickly to get information to support customers is important," Luke says. We are also developing AI-enabled tools to accelerate quote generation, which is a critical touchpoint in our customer journey.

Michael cites another example from Growth Hub, our innovation engine designed to solve industry-wide customer problems. "We’re exploring how AI agents could coach employees through the Growth Hub process, for example by constructively challenging ideas and helping teams assess how well they align with customer needs.”

To support productivity, we use Microsoft Copilot across the workforce. "Every employee has access to Copilot and AI training to assist with their productivity. We also invest in specific applications to support our manufacturing and operations, like using AI-powered visual inspection tools to ensure the highest and most consistent product quality for our customers." Luke explains.

The aftermarket opportunity

Our approach to aftermarket is at the heart of our customer-centric lifecycle support approach. Analysing, anticipating and servicing our customers’ needs is fundamental to our operating model, accounting for nearly half of Group revenues. We currently use data and machine learning in various ways to anticipate and respond to client needs more efficiently.

Data analysis helps map our install base and target aftermarket opportunities more effectively. "We have hundreds of thousands of valves globally," Luke says. "We can now use this data to target and generate sales for our existing customers more effectively, anticipating their needs with greater accuracy."

One example of innovative data use is the sensor technology we use to monitor valve performance in real time. "You can analyse the performance of a valve to make sure it is opening correctly. You can listen to the noise of it using a sensor to make sure it is opening and closing as effectively as it always has." Luke explains.

This approach to aftermarket growth means sophisticated data analysis can support IMI's strategic priorities without requiring a radical transformation of the business model.

Looking ahead

Both Luke and Michael are cautious about predicting AI's long-term impact, given the current fast pace of change. What is certain is that the technology is already transforming how we work and the impact we can have for customers.

"Engineering development cycles are shortening as you can use AI to test designs virtually," Luke explains. "You can imagine how fast that accelerates the customer's development cycle."

Another area where AI can really have an impact is predictive maintenance. "One of our customers' biggest frustration is their total cost of ownership," Luke explains. "If we can help them reduce that through more efficient predictive maintenance using AI and data, that is powerful."

Our perspectives are grounded in operational reality. Michael explains, "Advanced analytics techniques have been integrated into our business for years and have created a lot of impact. Generative and agentic AI technologies are very promising, but we are still at an early stage of adoption.”