Predictive maintenance is transforming manufacturing by harnessing AI and data analytics to pre-empt equipment failures, streamline operations, and drive strategic decision-making. Mark Venables explores how companies like McLaren Racing are using predictive insights to gain a competitive edge while addressing the challenges of data integration, security, and upskilling.
Predictive maintenance is redefining manufacturing, offering a proactive, data-driven approach to preventing equipment failures before they disrupt production. Through continuous monitoring and advanced analytics, predictive maintenance helps manufacturers determine when maintenance is truly needed, rather than relying on outdated schedules or reacting to breakdowns after they occur. The result is significant savings in time and cost, enhanced safety, and extended asset life – making it one of AI’s most powerful applications in the sector.
Evolving maintenance practices through AI
However, predictive maintenance is not new. “We have been living with predictive maintenance for well over a decade,” explains Alan Jacobson, Chief Data and Analytics Officer at Alteryx. “Most of us experience it in our daily lives. Take cars, for example. That oil light is not just a random feature; it is a simple form of predictive maintenance. Initially, it relied solely on mileage, triggering every 5,000 miles. Over time, these sensors have grown more sophisticated, factoring in environmental conditions like temperature and engine load. In manufacturing, the same concept has advanced, with algorithms and models that are increasingly capable of analysing vast amounts of data.”
Predictive maintenance in manufacturing has evolved from basic statistical models to sophisticated AI-driven insights, using varied and complex data sources. The type of data feeding into predictive models has grown significantly in scope. “Manufacturers are now able to bring in signals like infrared data to assess temperatures, X-ray and visual data for equipment analysis, and sound signatures – signals that ten or fifteen years ago would have been unthinkable,” Jacobson points out. “It is no longer just a single metric but a complex combination of information that these models can digest to provide actionable insights.” This combination of data sources offers a much fuller picture, giving manufacturers the information needed to anticipate and address maintenance requirements more accurately.
The role of data integration in predictive maintenance
The capabilities of tools like the Alteryx Analytics Automation Platform are central to advancing predictive maintenance practices. Today’s manufacturing environments generate enormous amounts of data from IoT sensors and machine outputs, yet much of this information remains siloed. “You might have equipment with its own predictive maintenance functions, but the real power comes when you bring in additional data from the surrounding environment,” Jacobson says. Alteryx’s tools enable manufacturers to pool information from multiple sources, combining internal machine data with external variables such as air quality and humidity, to produce a more comprehensive predictive model. “I often hear about companies building AI directly into their equipment, but no single piece of equipment contains all the data needed to create robust insights.”
The promise of predictive maintenance extends beyond equipment to the manufacturing workforce itself, enhancing productivity across all operations. “Imagine a large assembly facility with thousands of workers who all need to be at their stations by 7am. Naturally, not everyone shows up on time,” Jacobson elaborates. “Supervisors often spend 10 to 15 minutes working out which workers are missing, locating replacements, and getting everyone in place. Using badge scan data from employees entering the building, the Alteryx platform automates this process, notifying supervisors before the shift starts about absent workers so they can assign replacements immediately. This is not about complex algorithms but about applying data to solve real, everyday challenges in the workplace.” Such applications of predictive analytics demonstrate its potential beyond traditional equipment maintenance, highlighting its role in optimising workforce management and reducing stress for frontline staff.
Handling big data and IoT complexity
At the heart of predictive maintenance is data, but it is the volume, variety, and processing speed of this data that truly unlock the value. IoT sensors, though essential, bring challenges of their own. They produce massive datasets that require advanced tools for processing. “IoT sensors have been around for some time, but they bring two major challenges,” Jacobson notes. “First, they generate data in vast volumes. Trying to analyse all of it on traditional tools is simply impractical. Alteryx’s high-speed computational capabilities are crucial in this space, handling millions, if not billions, of data rows and integrating multiple data formats to deliver real-time insights.”
The second challenge is in making this data accessible to non-technical staff, empowering them to make data-driven decisions without coding knowledge. Alteryx achieves this through a drag-and-drop interface that enables engineers and operators to interact directly with the data. “Gone are the days when a manufacturing engineer would write a spec, hand it over to IT, and wait for results,” Jacobson remarks. “Today, domain experts need to work directly with the data to create meaningful insights.”
Security is also paramount in predictive maintenance. Manufacturing data is sensitive, with implications for compliance, intellectual property, and operational security. Alteryx supports highly regulated clients in fields such as healthcare, defence, and financial services, offering flexible deployment options to secure data appropriately. “We follow the latest standards in encryption and authentication, allowing organisations to protect their data while maximising its value,” explains Jacobson. “Interestingly, the greater risk often lies in not using data. For example, think about security cameras in a facility; if they are not monitored, or if data from these cameras is not analysed, you could miss critical incidents. The risk of missed insights from unused data is, in my view, becoming greater than the risk of using the data itself.”
Upskilling to address talent shortages in AI
While the technical challenges of predictive maintenance are considerable, a significant obstacle is the industry-wide talent shortage. The skills needed to implement and manage AI-powered predictive maintenance are scarce, and many manufacturers lack the resources to hire data science teams. This issue is tackled head-on through extensive upskilling initiatives, offering training, certification, and hackathons to build data competencies across organisations. “I am passionate about this because upskilling changes lives and careers,” Jacobson explains. “When you teach someone, like a manufacturing engineer, to work effectively with data, it benefits the entire organisation. The only real solution to the talent shortage is education and upskilling, which is why we prioritise this so heavily.”
McLaren Racing’s predictive maintenance strategy
A real-world example of predictive maintenance in action is McLaren Racing, where the power of data analytics plays a critical role in enhancing both performance and efficiency. Each Formula 1 race weekend generates around 1.5 TB of data from sources such as telemetry sensors, computational fluid dynamics, and the wind tunnel. With Alteryx, McLaren Racing processes this data in real time, allowing engineers to optimise vehicle design and race strategy at each stage. Predictive maintenance is integral to their approach, helping McLaren manage everything from component performance to driver telemetry, supporting the team’s competitive edge.
McLaren’s use of predictive maintenance extends beyond the racetrack. Alteryx enables the team to integrate data across different business functions, such as marketing and finance. For the marketing team, Alteryx’s geospatial capabilities allow them to correlate fan data with location information to create targeted engagement opportunities. The finance team, on the other hand, uses predictive analytics to align with commercial goals and ensure compliance with Formula 1’s budget caps. This cross-functional approach demonstrates how McLaren leverages predictive maintenance not just to improve performance but also to optimise the entire organisation, staying agile and data-driven in a high-stakes environment.
Predictive maintenance stands at the forefront of manufacturing’s digital transformation. The journey may be complex, but the benefits are undeniable: cost savings, efficiency gains, and strategic insights that help organisations stay competitive. For manufacturing leaders exploring predictive maintenance, Jacobson recommends a collaborative approach. “Bringing in insights from technology providers or peers who are further along in the journey can be invaluable,” he says. “Learning from others makes these projects less daunting, and many are surprised by how quickly people can gain data skills. It is not only about training your team; executives need to understand these tools too. I recently led a session with senior leaders on AI and LLMs to help them understand how the technology works. When executives are confident with the tools, they can make informed decisions and guide their teams effectively.”
The future of predictive maintenance is promising, as more companies embrace AI-driven data practices. Manufacturers investing in robust analytics platforms, upskilling their teams, and balancing security concerns will find themselves well-positioned to benefit from predictive maintenance’s full potential. By tackling the challenges of data integration, accessibility, and security, companies can leverage predictive maintenance as a competitive advantage, making it an essential part of the data-driven future of manufacturing.
