Estelle, welcome. Great to see you. Thanks very much for being with us. Let's just tell our audience a little bit more about what you do and your company does. So Estelle leads the global reference in ecological transformation. 215,000 employees across five continents, pioneering innovative solutions that combine water, waste, and energy management. And you serve over 111 million people with drinking water daily. It is huge and a huge responsibility. Um, Estelle, we know that, or you're going to tell us a bit more about how AI is integral to the work that you do in your business, but tell us about your personal AI awakening.
Oh, I guess, you know, like AI started with me with, of course, reading a few articles and a bit of science and, you know, things which are supposed to be intelligent. I was like, okay, you know, why not? But that was it. The real clicking moment for me was actually my daughter, who at that time was 17 years old. And I bumped into her bedroom and she was using ChatGPT—not to mention any name today—and she was supposed to have something important, you know, as a test on the next day. And I was like, okay, what are you doing, you know, using ChatGPT as opposed to rehearsing? But she was doing exactly the most intelligent thing I would never have thought of. She put all the lessons in the ChatGPT system and asking the AI to ask her questions back so she can rehearse her lesson. So the machine was asking her questions back and correcting about the answers. And I thought, wow, that's exactly how AI should be. And therefore, it's much more powerful than just a nice-to-have here, because it can enhance your learning in that case and many more, you know, opportunities even for a company like Veolia.
That's amazing. I'm just struggling to get my 12-year-old daughter to not use ChatGPT to answer all of her homework. I think your daughter's a lot more clever. So tell us now a little bit about the AI transformation in Veolia and with so many employees.
So the AI transformation, or GenAI as opposed to the classical machine learning, I guess, you know, is on the way. So we are learning, you know, every single day, new usage and new opportunities and new risk as well. So the main thing for me is to use AI to help Veolia be even more successful in what it does, which is to reduce the water footprint of industries, to produce more local energy and green energy, and to reduce pollution. This is what we do as Veolia. So the first question we've asked ourselves is, how can AI help us in this journey? And we've developed a series of tools which already are, you know, starting the scale-up moment. We are done with a proof of concept and MVP. We're really more on the ramping up already. One which I think is amazing is Talk to My Plants. Talk to My Plants is using the natural language model for a guy in the middle of the night in one of our plants—one of our thousands of plants across the globe—to help him or her with the maintenance and making the plant more efficient, again producing more energy, more local energy. So the idea is to use the democratization power of GenAI because it's a natural language model. So the technician in the middle of the night, you know, would ask if there is a turbine which is a little bit less, you know, efficient and he doesn't know what's happening, he would ask his cell phone in language model as in, what's happening to the turbine? Can you tell me what the source of the problem is? And the machine, of course, using live data together with, you know, digital twins together with his deck, will answer him: it looks like this valve here, you know, is having a little bit of pressure problem. Maybe it's not that, but you should have a look at the valve in priority. Saves enormous amount of time and in the end it means more energy produced, more green energy produced for exactly the same outcome. So in this example, what I think is absolutely fascinating about GenAI is the democratization of it. As in, any single technician pretty much can use it, like every single person on earth can use, you know, on their cell phone the ChatGPT example. So we are doing that already for wastewater treatment plants, for desalination plants, for energy from waste and so on and so forth. And the amount of, I guess, environmental benefit we can drive from it is absolutely enormous.
How have you found the adoption of AI within your organization and so many employees? Because we've heard a lot over the last two days about trust and some people perhaps struggling with that a bit more.
I guess, I guess trust cannot be something you decide. You have to build it milestone after milestone. So we started our journey pretty much two years ago and we didn't know what we didn't know. So, you know, instead of saying, okay, the guys at the top of the company do know and you, you know, average employee, have just to listen, was not at all the way we've done with AI. You know, we've gone through the journey of, you know, embarking, including the union representatives, for instance, who were worried about potentially job losses, and say, okay, we're going to learn together, all that. And we've trained progressively 13,000 employees who are, you know, AI ambassadors, and they are trained and they are at pretty much every single level in the organization, managers and non-managers, exactly in the same way. Again, I think the democratization element of it is something which is super powerful compared to the classical digital where you need data scientists which are super specialist. Or something here, you can precisely, you know, empower 215,000 employees.
As we all know, AI and data centers consume a huge amount of resources, of energy. How do you reconcile that?
It was at the forefront of our choices, including in terms of provider, from the start. As you can imagine, when you're Veolia, you don't want to use more resource via the data center than the resource you would save by being more efficient in what you do. So the costs have to be lower than the benefit, including environmentally speaking, and not only in terms of euros. That was at the start, you know, very much, you know, at the top of our agenda. And in addition to that, we have developed an offer for data centers to be more sustainable across the globe. And we already have 100 references across the globe, in the US, in Europe and elsewhere. Where we recoup the wasted heat, we recycle the cooling water, and we can recycle or treat as well the electronic waste which goes with the data center. So we are making it so that, you know, it's a sustainable data center offer of our services in Veolia.
We've been talking about how many people you employ. When you look at the use of AI, is there a question about whether it will replace some of those people?
Of course, the question was asked from the start by our employee and everybody. But the answer is no, and we're pretty certain about that. The answer is no. It doesn't replace human beings. It enhances human beings. It helps them save boring time and repetitive tasks to be able, therefore, to be focusing more on what creates value. And it goes all across the entire chain, you know, across the entire organization. And as I said, you know, we trained a lot of people because the transition needs to be a bit, you know, organized. We are training, you know, a lot of people to be able to go beyond the first impression of, oh, what's that, you know, and the kind of potential antagonism. And I think, you know, it's a very powerful tool. And I can say we've put the human beings and our resources, which are the employees of Veolia, at the center of the revolution, as opposed to, you know, on the side.
What do you think are the biggest challenges of adopting AI at the moment?
