(AFP / JOEL SAGET)
“Does it cost more energy than it brings in?”: the French Schneider Electric, giant of electrical equipment and industrial automation, is increasingly relying on artificial intelligence (AI) to help its customers to consume less energy.
The group wants AI to meet a “need” of its customers, explains Philippe Rambach, first “Chief AI (artificial intelligence) officer” of the group to AFP.
“Very often, people start from technology, say ‘here, in AI, we can do that, how could we use it?’ We reversed the subject, we said, what are the issues business of the company and in that, which ones can AI help?”.
Among the concrete applications developed by the group, one has the sweet name of “Microgrid Advisor”.
This software, assisted by sensors, “will be able to predict the consumption of your building within 48 hours (updated every quarter of an hour), to predict the production of your solar panels, your batteries, the quality of the network for the next 48 hours and after to optimize”.
It will then allow the user to use the current produced by their solar panels, or to store it, to buy from the network, to sell to the network, explains Mr. Rambach.
An optimization which can make it possible to further green the use of energy: “when electricity is expensive, it is also carbon-intensive, because that corresponds to the moment when everyone wants electricity at the same time, and in these cases- there, the supplier is obliged to start the coal and gas power plants,” underlines Mr. Rambach.
This system, which was originally used to protect against network losses, perfected using AI, is currently reserved for professional customers. It will be available to the general public “before summer” under the name “Wiser”, according to Mr. Rambach.
Another tool developed this time for internal use: software that automatically recommends to sellers what they should sell to a customer of this or that type.
The tool, based on the two million quotes stored by the group over five years, allows sellers “on the one hand to save a little time, but above all to think about everything they could sell”, explains Mr. Rambach.
These two applications use a classic AI algorithm, such as “machine learning”.
– “Return on carbon investment” –
Overall, the group uses little generative AI, which is recent and very energy-intensive, explains Mr. Rambach, for whom, beyond the financial return on investment, “we must look at a carbon return on investment”.
Internally, it is used mainly to write code: “we are going to deploy a Microsoft solution, called Github Copilot, which helps our teams write code faster and of better quality,” explains Mr. Rambach.
Another application of generative AI: making customer service easier. When the customer calls to ask a question, “we are able to construct an answer” for the operator. The latter, informed of the source of the response, “will check in the document (proposed by the AI) that the response appears correct to him and if this is the case, he validates it and transmits it to the client”.
Asked about the energy expenditure that AI could generate in the broad sense, while the group’s main purpose is to provide solutions for optimizing consumption, Mr. Rambach is very confident: “for the moment, we are really on a ratio of one to 100 or one to 1,000, when we see what we spend in relation to what we save.
Regarding possible fears of corporate restructuring brought about by the use of AI, he is also optimistic.
“AI never replaces an employee. AI is something quite precise that does a task and fortunately, in our company, there are not many single-task employees,” explains Mr. Rambach.
On the other hand, according to him, AI helps the group to remedy its recruitment difficulties, particularly for maintenance agents who repair equipment: “We use AI to help them do their work more quickly, more easily , to be able to serve more customers.
The teams dedicated to AI have seen their workforce jump, increasing in two years from 40 to 350 employees.