The Role of AI in Asset Performance Management

What should asset management practitioners realistically expect from the overarching potential of artificial intelligence? The market is teeming with solutions that make big promises about an organization’s ability to predict, optimize and automate. Generally, the topic of AI comes with a lot of fear that stems from a lack of knowledge of the applicability of AI and its limitations. Let’s first go down one level to learn about the different types of applicable forms of artificial intelligence to help us learn about what these tools are capable of today and how they can be used to enable us, as humans, to do more of what we are good at.

Sensor-Based Analytics

An article by TIME featured a story about how the Apple Watch helped a gentleman to seek medical care in good time.

“[Kevin] Foley, a 46-year-old information-technology worker from Kyle, Texas, was heading into the theater to see Avengers: Infinity War when he realized he was having trouble breathing normally. The sensation struck again during another movie the following night, but more severe this time. Once the credits on the second film rolled, Foley took action.”

Following notifications from his watch that an irregular heartbeat was detected, Mr. Foley got himself to the ER where he was treated for atrial fibrillation. The Apple Watch he had been wearing contained special medical sensors that were being tested by Apple and Standford University’s Medical School.

Applying AI to the Data You Have

Today, in the asset performance management market, sensor-based analytics are extremely popular. Sensors are not new, of course, but today they are associated with the topic of AI. The type of machine learning models that are applied here produce classifications, predictions or decisions about the input data coming from the sensors. These models need input data (sensor data), training data and feedback data to make predictions. One of the most innovative ways to put this power into the hands of asset managers is to “leverage AI to analyze large volumes of operational and condition data to get in front of unexpected equipment failures.”

This is the type of AI that Mr. Foley benefited from with his smart watch. The watch’s notification may have saved his life, but is this the way we want to manage our own health? Are we really going to wear these devices to alert us only when some serious or severe condition is detected? The same reasoning applies to the sensor-based predictions regarding the health and condition of assets. When AI is used in asset performance management (APM), it can enable an organization to get - and stay - ahead of failures and vulnerabilities rather than simply give them a way to react to those threats sooner.

Generative AI

Generative AI or GenAI has been popularized by the emergence of ChatGPT and other Large Language Models (LLM). This is the domain in which most of the big advancements in AI have taken place and have attracted the most attention. Websites and whitepapers describe how top management intends to invest in GenAI. So how can ChatGPT or Gemini be applied to industrial processes? If so, what criteria exists to measure how well these LLMs should perform in industrial use cases?

In this series on AI in asset performance management, you can look forward to a meaningful discussion on AI capabilities. We’ll explore why “almost all CEOs (99%) are making or planning significant investments in GenAI.” While it’s not as well known as generative AI, we’ll discuss the applicability of discriminative AI and the impact it can have in the performance of your asset management system. This will add up to some useful guidance on how AI can help you achieve asset management system performance that may otherwise not be possible.