Key points
AI in Industrial Asset Management
Keith Smith, COO at IntelliAM, shares his insights into the role of artificial intelligence (AI) and machine learning (ML) in asset performance management within the modern-day manufacturing industry.
In today’s manufacturing environment, it’s no secret that traditional predictive maintenance (PdM) methods – such as vibration analysis – are struggling to keep up with increasingly complex machinery and diverse drive technologies. Not only is this frustrating for process industry professionals, but it is significantly impacting productivity potential across the sector.
That’s because in the past, assets used to operate at fixed speeds and drive systems were simpler, making the deployment and analysis of traditional PdM tools far easier to apply. This meant that vibration was often a reliable indicator of potential machine failure.
But with the advent of multiple drive technologies, variable speeds and loads, and a wider range of failure modes, relying on vibration alone can paint an incomplete, or even misleading, picture of asset health.
At the same time, manufacturers are under mounting pressure to do more with less.
Their goal is to keep production running, costs down, and every asset on the factory floor in peak condition. But achieving this against a backdrop of rising demand, shrinking budgets, and stretched resources, is no easy feat.
When the tools designed to support manufacturers in running their factories more efficiently are instead hindering it, this makes progress difficult.
A combination of greater false positives, late warnings, and missed issues is often the result of applying outdated methods in an industry that has evolved and moved forward. This creates an environment where unexpected breakdowns and downtime are all too common, threatening both production and operational progress, as a result.
Another widespread challenge is the ability to gather the right data at the right time. In sectors such as food and drink manufacturing, equipment often operates in stop-start cycles, meaning many IoT sensors can’t detect when the most meaningful and relevant data should be captured. This is what makes continuous monitoring essential.
“A mindset shift is needed, from one that views AI as a bolt-on tool or a threat to jobs toward one which recognises it as a strategic capability.”
But continuous monitoring alone isn’t enough. To generate meaningful insights, contextualisation of what the data really means is key.
Accounting for operational variables – including loads, environmental conditions or drive behaviour – is what enables data to be interpreted in a meaningful way and one that drives true value for the business.
The same principle applies to productivity measurement tools like Overall Equipment Effectiveness (OEE), too. While traditional OEE systems report machine data and highlight downtime trends, they still rely on manual interpretation of all that data.
The next step involves taking all underlying PLC data and system signals and using machine learning models to identify patterns in alarms, warnings, and operating behaviour – pinpointing root causes without the guesswork.
Even with large volumes of data at their fingertips, many manufacturers still face another key challenge – data unification. It’s common for a single machine to have data spread across three or more separate systems, each one using its own unique terminology and structures. Without a way to consolidate that information, forming a complete operational picture of asset performance is nearly impossible.
That’s why a platform that’s built for unified data is central to the modern-day manufacturing tech stack. Without this, data can’t be tied together for insights, meaning valuable context is lost, patterns are missed, and maintenance decisions remain reactive instead of predictive.
Manufacturing assets require a holistic, context-aware approach to unlock predictive insight.
Outdated predictive maintenance strategies are keeping many manufacturers locked in a costly, reactive loop. That’s because without investment in the right asset management strategy, where predictive maintenance plays a central role, manufacturers are left either responding to failures as they happen or performing unnecessary, intrusive maintenance that can cause more harm than good.
It’s the equivalent of dismantling a car engine every year ‘just in case,’ even when it’s running smoothly. This often leads to reduced performance and unnecessary cost. It doesn’t make operational or financial sense, especially in complex manufacturing environments.
A more innovative, data-driven approach allows manufacturers to take action only when needed. To put this into context, instead of sending a functional pump away for an expensive overhaul every year, engineering teams can now monitor specific indicators like bearing performance and apply targeted interventions, such as lubrication, before any issue arises.
This approach not only reduces unplanned downtime but also avoids unnecessary maintenance spend and eliminates the need for Original Equipment Manufacturer (OEM) involvement in annual asset checks.
In industry, maintenance is widely seen as a cost, but with the right strategy and technology in place, it can shift from being a cost centre to a value driver that actively pays the business back.
AI and ML are redefining the way manufacturers approach asset management. These technologies are providing a shift away from reactive maintenance to more a more proactive strategy that’s backed by data insights.
That’s because by tapping into billions of data points across machines, manufacturing teams can begin to identify patterns, reduce waste, increase throughput and efficiency, and improve operational decision-making.
“AI and ML are redefining the way manufacturers approach asset management.”
However, the move to intelligent asset management isn’t a one-step solution. It’s a journey that’s most effective when taken in logical stages, starting with understanding asset criticality and redesigning maintenance strategies to replace outdated, intrusive routines.
The next phases involve defining how and where to capture data and then layering reliability data with operational context to create more meaningful insights. And as a manufacturer’s AI maturity grows, productivity gains can be unlocked through self-learning machine learning models that optimise flow rates, speeds or other parameters in real time – boosting OEE, reducing waste, and improving energy consumption.
When adopted and applied in this way, AI can support engineering teams to make smarter decisions, reduce unplanned downtime, and extend asset lifecycles – transforming asset performance management from a costly burden into a strategic advantage.
Ultimately, hidden inside the plant and machinery of every factory in the world lies a largely untapped opportunity, and one that doesn’t necessarily require new equipment but a new mindset.
That’s because the challenge isn’t data scarcity, but how to unlock its value.
PLCs, IoT sensors, and cloud systems are already generating a rich stream of data, from oil temperature to energy consumption and machine alarms, it’s simply a case of implementing the tools to access those untapped insights.
“Ultimately, hidden inside the plant and machinery of every factory in the world lies a largely untapped opportunity.”
The reality is that AI has the power to transform how people and machines interact, but while the potential for seismic improvement is real, this is only possible if manufacturers embrace the evolving nature of predictive maintenance.
A mindset shift is needed, from one that views AI as a bolt-on tool or a threat to jobs toward one which recognises it as a strategic capability that thrives alongside human expertise.
By training and refining AI models, engineers enable systems to become self-learning while embedding AI into daily operations. This creates new roles focused on integrating AI into manufacturing workflows.
AI is most effective when coupled with manufacturing and engineering domain expertise. This is because it needs to be taught, tagged, and refined by those who understand the machines it monitors. This not only enables the system to become self-learning, but it also creates new jobs that embed AI into the heart of industrial operations and decision-making.
Ultimately, a new industrial revolution is here, and those who take this approach stand to gain more than just cost savings. They will be able to unlock both resilience and a competitive advantage that’s built on foresight rather than hindsight
What role does AI play in industrial asset management?
AI enables predictive insights by analysing large volumes of operational data to detect anomalies, predict failures and optimise asset performance
Why are traditional predictive maintenance methods no longer enough?
They rely on single indicators such as vibration analysis which do not account for variable speeds, diverse drive technologies and complex failure modes
How does AI improve predictive maintenance in manufacturing?
AI and ML provide contextualised insights by unifying sensor, PLC and operational data to identify root causes and recommend targeted actions
What are the main benefits of AI-driven asset management?
Benefits include reduced unplanned downtime, lower maintenance costs, better OEE scores, optimised energy use and longer equipment lifecycles
Can AI replace engineers in asset management?
No, AI is most effective when combined with domain expertise because engineers teach, tag and refine models to ensure accurate monitoring and decision-making
What is the first step in adopting AI for asset management?
Start by understanding asset criticality and redesigning maintenance strategies before scaling with unified data platforms and machine learning models
The case of former referee David Coote’s out-of-work comments, made towards the end of last…
Since being asked to write a column I have been very careful to not drift…
I’ve written quite a lot about problems with the use of AI chatbots, modelling and…
TL;DR Summary Box Smarter Power Infrastructure for Net-Zero Process industries face pressure to decarbonise and…
TL;DR Summary The UK chemical supply chain underpins 97% of manufactured goods and faces pressure…
TL;DR Summary Water recovery helps food and drink manufacturers reduce water costs, meet discharge rules…