Accelerating Artificial Intelligence in Asset Management – Part I
Have you watched “Formula One Drive to Survive”?
This Netflix series showcases the amazing engineering effort behind the top performing racecars in the world. But why are we discussing F1 racecars in a blog about AI (Artificial Intelligence)? This Netflix series tells a story about decisions, how they affect outcomes for every race, and how the use of data and information affects the quality of those decisions.
The current state of AI is like Formula One racecars. These machines depend on advanced technology and specialized skills. They require a highly technical engineering team coordinating their efforts to ensure the racecar goes around each lap of the track, in the least amount of time possible. This is a process of continually compounding the team’s collective understanding of their cars, the tracks they race on, their competitors, and the conditions they need to adjust for before and during the event.
Harnessing the power of AI is similar to harnessing the power of an F1 race car. The team of engineers, subject matter experts, and data scientists precisely coordinate their efforts to use information they have (historical data) to produce information they don’t have (predictions). The quality of these predictions directly enables the quality of decisions made by heads of the team. At this elite level, the aggregation of marginal improvements is what makes the difference between a podium finish or not scoring any points at all. This elite and prestigious level of performance can attract and retain the top available talent to be part of this team.
Similarly, to harness the power of AI, a team of data scientists, subject matter experts, and specialized software engineers are required. This means that only those organizations that can attract and dedicate such a team are going to reap the rewards promised by AI. At least for now.
When you consider the application of AI within industrial facilities, building a team to drive successful AI projects can be even more challenging. A skills gap due to an aging and retiring workforce, the need to collaborate across many organizational functions such as operations, maintenance and IT can create significant burden on already stretched resources. One of the keys to being able to successfully leverage AI at scale, is to put the power directly into the hands of equipment and process experts. Yes, AI as an enabling technology provides great productivity gains by rapidly assessing large volumes of data on a continuous basis, but the true value drivers are not algorithms. It’s the people and their knowledge of how the production process operates, how equipment can fail, the symptoms that indicate potential failure and most importantly what action to take when anomalies are detected.
In the APM space, we have seen AI applied to the most critical and complex assets, yet these represent a very small percentage of the total assets in an industrial facility. If we could turbo charge the process by enabling equipment experts to directly build anomaly detection models at a faster rate, we can unlock much more value across a broader set of challenges related to the risk, cost, and performance of assets.
It wasn’t the sophisticated engineering that went into the car that brought about the era of motorized transportation. It was Henry Ford’s assembly line production which brought the car to the masses. AI is going to continue to advance and will do so at an increasing rate. But until we solve the challenges of practicality and availability, we will be dependent on precision-coordinated teams of highly skilled experts.
Look out for our next post where we will offer some insights into approaches to fully harness the power of AI.