New AI technologies offer ways for companies to automate and optimise many processes, French multinational Schneider Electric told a media briefing last Friday (7 March). The tools can support and sometimes even replace human judgements in industrial operations, said digital transformation lead François Disch – but major “hygiene” efforts are needed before that can happen, he said.
Product-centric companies with highly evolved supply chains are likely to have mature data, Disch said, such as tech behemoths and automotive companies. But they are just a small part of the industry, with “most, if not all” small to medium companies operating with immature data management practices. The issue is global, Disch said, but the UK is “quite low on average”.
Companies need to structure and organise their data so it accurately reflects their processes and manufacturing methods before they think about AI, he said. The same is true in energy, he continued, where there are a lot of use cases for AI, but not enough infrastructure to collect and manage data.
“When the data is clean, we’ll be able to apply AI. Dropping it in the middle of the way is not necessarily going to help much,” Disch said. “There are different ways of applying AI; not everything is responsible.”
If the problems can be solved, the company said the opportunities are many and varied. AI is the next industrial revolution and can support two of the government’s key priorities, said Maria Tjader, Schneider Electric head of government affairs for UK and Ireland: Clean Power 2030 and the upcoming Industrial Strategy.
“If you deploy AI in the energy grids, you're going to be able to monitor and optimise energy consumption, energy use, ultimately leading to being able to… optimise and improve energy efficiency. It's also going to help to integrate renewable sources better, and also, I think, critically manage balancing the grid,” Tjader said.
“On the industrial strategy side, industries can use AI to optimise, plan and predict demand and output. It also can be used to manage stock and supply chain, and also – both in industry and in the energy system – to be able to forecast energy demand and production.”
Specific industrial applications include estimating the remaining lifetime of assets, waste reduction and downtime prevention through anomaly detection and risk estimation. In the energy sector, AI tools can tackle tasks including forecasting renewable production, process planning and smart electrification of heat management.
The energy demand of AI technologies themselves has received a lot of attention, which will likely increase as net zero targets draw closer. Widespread inefficient deployment could lead to a potential energy crisis, according to Schneider Electric forecasts. It called instead for efficient digital infrastructure and robust governance to enable sustainable AI use.
Not all companies need to use large models, Disch said, and many could use smaller ones that are more efficient. The first step is to get a specialist who can “drive you to the right solution,” he told Professional Engineering. “All the different solutions have a very specific type of application… there are several ways of skinning the cat.
“Going for a human-like way of solving the problem is going to cost you the most energy, while if you’re looking at other ways, thinking a little bit outside of the box, you may be able to find cheaper, more targeted ways of finding solutions. And if you've got an AI specialist, together with an operational specialist, you'll be able to drive to a solution that is the most efficient in terms of energy or computation effort, as well as skill. It’s not just replacing a human with a robot.”
Natural language processing techniques such as text classification have relatively low associated emissions, for example, while computer vision techniques such as object detection are higher. Generative AI – most often associated with chatbots such as ChatGPT – have even higher associated emissions, according to a 2024 study from AI company Hugging Face and Carnegie Mellon University, with image generation potentially hundreds of times worse for the environment than text classification.
Schneider Electric established an AI-focused unit after recognising the importance of the technology in 2016. It now has more than 350 specialists, working at international hubs in the US, China, France and India.
Want the best engineering stories delivered straight to your inbox? The Professional Engineering newsletter gives you vital updates on the most cutting-edge engineering and exciting new job opportunities. To sign up, click here.
Content published by Professional Engineering does not necessarily represent the views of the Institution of Mechanical Engineers.