For years, digital transformation seemed to stagnate. Operational technology (OT) teams were implementing new sensing devices and software, and collecting massive amounts of data, but they had trouble turning that data into actionable information. In some cases, that was a result of new silos created in their organizations by the very technologies designed to bring about digital transformation. In others, it was because the teams had access to loads of data but didn’t have the personnel or expertise to do much with that data once they had it.
However, in the last three years, a new technology – artificial intelligence (AI) – has come to center stage to consume and transform mountains of data to drive improved operational insight. Industrial AI, Amy Schantz explains in a recent article in Industrial Ethernet magazine, will bring new solutions for using data captured as part of digital transformation initiatives to drive operational excellence.
Industrial AI at the edge
It all starts at the edge, Amy explains.
“Connected to AI is evolving and increased focus on edge computing. Edge computing brings AI resources close to the source of reliability issues to track and trend equipment health, helping personnel make critical decisions in real time.”
This has a couple significant implications for OT teams. The first is that the AI models used in OT are likely to look different from the generic large language models the public is most familiar with.
“Much of the future of AI is unknown territory, and the potential for it to create new challenges can be unnerving, especially as we consider AI’s use in real-time, mission-critical applications. However, what we do know is that enabling intelligent automation with the help of AI copilots and agents is going to be central to both user experiences and increased productivity.”
That focus on real-time, low-latency, mission-critical operation means that industrial AI software will need to be deployed where the action is—at the industrial edge rather than in the cloud. Teams will rely more heavily on AI applications deployed on prem, which means they will need to be building toward the AI-enabled enterprise operations platform architecture that will enable seamless integration to move critical data, with context, wherever it needs to be consumed.
Industrial AI in action
Amy shares an example in the article,
“There is AI at the edge quality inspection—an example that we’re implementing at Emerson facilities. Using a next generation PACSystems™ IPC platform combined with PACEdge™, users can quickly and easily pass or fail equipment in quality inspection using AI tools and real-time inference. This solution can be scaled but remains air gapped in the factory, providing a low latency solution while allowing security to remain paramount. Edge analytics are also being deployed to deliver predictive maintenance for rotating equipment. Connected worker technologies are driving operational excellence, with AI copilots providing troubleshooting assistance in both operations and maintenance, and augmented reality overlays supporting decisions in the field.”
But to accomplish this, Amy explains, OT teams need to break down their existing data silos. The unified data fabric supporting a modern enterprise operations platform is designed to ensure secure, continuous access to data from the intelligent field, through the industrial edge, and into the cloud. That means that while the plant floor can be supported by well-informed Industrial AI applications running on prem to monitor and maintain real-time operations, the same data can also be simultaneously feeding the enterprise business platforms that can help companies make the best long-term decisions across a site, a fleet, or an entire enterprise.
“Unified, contextualized data will feed the more secure, next-generation automation technologies that provide fast response, are infinitely and intuitively scalable, and provide instantaneous decision support across every competency from the factory floor to the head office boardroom,” she added. “The result will be a transition to more autonomous operations, predictive rather than reactive maintenance, and faster engineering, development, and deployment cycles.”
Nobody can say for sure where the AI revolution will take industrial operations. However, a shift is coming, and implementing the foundational elements today that will simplify and streamline industrial AI adoption at the edge will help companies lock in competitive advantage in the years to come while also delivering safer, more reliable operations from the onset.
