Through combining large language models with generative AI and other types of applied AI, plants will radically transform their productivity and sustainability, writes Jim Chappell, Global Head of AI, AVEVA
Artificial intelligence (AI) is often pitched as a cure-all boost for businesses. But the reality is more nuanced. AI is not just one thing, and its many aptitudes are being developed over time in stages.
The term AI covers a wide base of technologies, including many types of machine learning such as predictive analytics, deep learning, reinforcement learning, and more recently, generative AI (genAI) with the massive new Large Language Models (LLMs).
I believe that AI technologies, once fully realized, will be more life-changing than the internet and have a profound effect on the way we run and maximize businesses. Plant systems will be transformed from start to finish. AI-based solutions could automatically optimize, strategize and execute operations – 24/7.
This reality is on its way but it is not here yet. In today’s world, we are seeing decades of AI research and testing come to the fore, combined with the humanization of AI through LLMs.
In this respect, AI-powered solutions are already driving already huge wins across industries.
These technologies are enabling organizations to reduce fuel consumption, improve carbon capture, lower emissions, maximize equipment lifecycles. They are boosting corporate sustainability while increasing efficiency and profits.
How are they doing this?
AI-driven software helps industrial businesses engineer faster through suggestive design and simulation; operate at optimal production levels for maximum efficiency; provide decision support to connected workers; reduce waste for maximum yield; and recognize potential equipment failures before they can occur.
Companies around the world are already realizing massive sustainability and productivity benefits through applied industrial AI. For example, US power giant Duke Energy saved over $250 million by installing predictive models which provided early detection of issues, reduced errors, enabled faster results and boosted ROI. Or take Schneider Electric, which was able to prevent three full factory outages by predicting issues in advance.
Infusing AI through the network
People often think of AI through a GenAI lens. However, GenAI is only one type of AI, along with many others, and the truth is many of these building blocks of AI have been in play for decades. However, as available data increases, the true value of AI is being realized. The new LLMs with GenAI are no exception.
Today’s experts are in the process of honing these mathematical algorithms and connecting the different strands of knowledge to provide the solution, or part of the solution, for industrial businesses.
I call it AI infusion. It’s a process of diffusing AI capabilities across the board in areas such as engineering design, simulation, operations, maintenance, value chain, and data. With the availability of the massive LLMs, we will begin to see a leap in operational effectiveness – when all relevant data, no matter how obscure, is made accessible and intuitive to the user.
Through the humanization of AI, users will get more value out of their data and eventually out of everything they do, including software functionality. It will likely become the new user experience for computers.
The next leap forward will come through integrating multiple types of AI and applications, so that workers can do their jobs to maximum capacity. Evolving technologies will see humans and AI technologies working together for maximum progress.
This development can already be seen in action in the form of digital twins – virtual representations of physical objects, systems or factories that are created through data gathered from Internet of Things (IoT) devices, advanced computer systems and digital processes.
AI is the brain behind the digital twin. Once a digital twin has been put into motion, AI can generate customized data analysis to support enhanced operations.
And by layering LLMs on top of the network, AI technology can become even more of a contemporary or assistant than a mere tool.
For example, by leveraging their company’s industrial data and a LLM, the user could ask natural language questions of the system with minimal setup required. The user might ask objective-driven questions, like ‘Which windfarm has the lowest energy output this week?’ or ‘Why is my compressor performing worse now than last week?”
By synthesizing GenAI and other types of AI with LLMs, users can access applied knowledge in fields that might be out of their ken, such as 3D design drawings and maintenance manuals. The LLM can extract the relevant information and use it to find issues and help solve problems – all from a prompt-based question from a human.
Eventually this means that workers and executives will do their jobs with maximum effectivity and ease, while operations run at optimal and sustainable levels.
In other words, this is what industry 5.0 will look like. AI-based solutions will run hybrid semantic searches, constantly scoring and finding relationships among different types of data. This information will be the key to solving organizational challenges on a massive scale.
As you can see, we are rapidly moving from narrow AI toward general AI where software will become much more human-like in capability. As the trend continues, AI will become more objective-driven, leveraging everything at its disposal to achieve its goal.
However, AI is not a replacement for human intellect, rather it should be viewed as an enabling partner. It is destined to progress from today’s software that assists humans, to software that works alongside humans – workers who are deeply connected to the technologies through easy-to-use interfaces.
One day, in the not-too-distant future, a plant manager might ask “Can you lower my factory’s carbon footprint? I give you permission to do anything you need to do that’s ethical and legal to lower my carbon footprint”. This command could then set off a chain of events including sending emails, looking at set points, checking assets for efficiency problems, making control system changes, and issuing new work orders if needed.
We’re not there yet. But we have most of the pieces, so watch this space.
About the Author
Jim Chappell is Global Head of AI at AVEVA. AVEVA is a global leader in industrial software, sparking ingenuity to drive responsible use of the world’s resources. The company’s secure industrial cloud platform and applications enable businesses to harness the power of their information and improve collaboration with customers, suppliers and partners. Over 20,000 enterprises in over 100 countries rely on AVEVA to help them deliver life’s essentials: safe and reliable energy, food, medicines, infrastructure and more. By connecting people with trusted information and AI-enriched insights, AVEVA enables teams to engineer efficiently and optimize operations, driving growth and sustainability.
