Artefact’s Post

Transcript

Hi. Hello, everyone. Thank you very much for being here. Yeah, here with us today. Thank you, Crystal. Yeah, we met two weeks ago actually during the special gathering. At least eight was a special event dedicated to French AI talents. And you were there, the speaker and you're still and you are here today to talk about. AI but So what makes a big company like Orange? You know, special and important player, if we can say in the in the second system, so we can say how did you adopt AI it? Well, I think I mean that orange, of course, and the given we we have millions, hundreds of millions of customers. We have networks, I mean thousands of kilometers of of networks and and obviously we have many employees, of course, like any company and you've seen the example of. I mean the car industry, the automotive industry. And so we've actually embarked on the AI transition. Of course, we AI wasn't that popular, the buzz was on something else, but we've been working on digital transformation. We used to call it that way, making sure we have the data, structured data in every part of the company. And I think just one thing because of course we want to bring it, we want to use AI to make our networks. Quarter to make our employees more efficient. And last but not least, and probably the most important to make sure we improve the experience of our customers. And you may remember that one year ago we talked a lot about the Olympics now, but there was the women World Cup and, and there there's one video that was actually we, we worked on it with a Marcel, the publicist agency to promote our commitment to women in sports. Everyone talks about Mbappe this these days. But the reality is that we are promoting diversity for sports. And so we asked our agency and Orange Force worked on this video that was combining AI, AI generated video to promote women. This video was not pushed through commercial on TV, but it was. So it created a buzz and we have 2 billion views on the video we've got. We had the many awards. So that shows that AI is is already embedded in many of the things we do in and obviously we work with. But the fact on the on many experiments so in our in our networks, I'm afraid some of you may. Have had the complex experience about the connectivity, fiber connectivity. And so we embed a lot of quality similar to what you saw for value, a lot of quality touchpoint, leveraging video, leveraging pictures to to make sure we can proactively anticipate failures in the network. In our network operation center, we have already experiments where by leveraging AI, we divide by three the number of alarms that are experts. Have to deal with of course in our I mean customer centers we embed AI to make sure that we can fine tune the relationship with the customer personalize the message and the exchange with customers, but also make the life of our center agents more efficient. And when you call a call when you call a a contact center customer center you know that you you waste some of your time listening you listen to it and you see the agent that's typing on the on the on the on the keyboard because. Simply they have to enter your information etcetera. So now by leveraging AI, we, we make the experience more, more, I mean easier. And so we save time for the customer, we save time for, for the agents. The one thing that of course we are working on and and that's a big question mark for the industry is that we talk about compute bottleneck, we talk about power bottleneck. The real question for us is how do we build the best networks for AI? And if you have seen the latest, I mean GPT for all when you have real time translation or real time interaction and actually. If you want to have real time experience like voice, you know the old voice that you don't use anymore. Internet networks have been built on best effort basis. But if you want to divide by 10 the latency of exchange when you talk to an AI agent, if you have a personal agent in your in your mobile phone. That requires a fundamental change in the way we designed the last mile of the network. So that's something that we are working on actively with our experts. And this is something that today we have we we're not sure yet what's going to be the killer app. If you have a fiber connectivity at home, there's no doubt that you'll have enough bandwidth and that the latency experience. But again, we've built our network today for video, for video streaming and for video exchange because the boost of video because you do video conference. Polls and you watch videos online. We've added cashes in our networks. We have content delivery networks. But AI is going to change fundamentally the way we'll have to design our networks. And that's something that we still don't know what's going to be the first adoption where how it took 10 years to adapt our network for video, a bit more than 10 years. So we know we are on a journey and the challenge which is power as I think Christoph was, was saying it. We will have to also adapt. Today we're using AI for a lot of proof of concepts. Without putting the, the consumption behind it, but the bottleneck will be such that we will have also to embed compute where it should be to minimize power consumption. And, and that's something that of course we're working on actively as an industry. It's not just orange, but of course, being a leader in Europe, we, we are taking the lead on that. And of course with AI, of course it's a great technology, but also it's it, it creases the carbon emissions as you. That said, and I'm sure you have, yeah, you have the ambition to be net 0 carbon by 2040. So I how are you going to manage that with all the things you're doing with AI? Yeah. So, so the, the industry is committed to reducing its carbon emission and we are committed by a 2014 net, net neutral. And, and for us, we consume electricity and, and, and in France, indeed the which is our largest country, we, we consume a lot of electricity. It's highly decentralized consumption because it's every mobile site, it's every point in our networks. But it's true that today the digital industry, it's 2.5% of the total carbon emission. The reality is that if we don't do it right. The risk is that it becomes 7% by 2040 of the total carbon emission. Telecom networks are very small part of that. So it's not just us. We have to work with all the big platforms on this. But for us, we know what we have to do. We have to obviously consume renewable energy. In France, we are lucky because we have decarbonized energy, but still we need to reduce and the challenge for us is we continue to have a traffic on our networks that's increasing by 30% every year. So for us, not increasing the consumption of electricity is already a massive challenge in a country like France, we managed to reduce by 10% the, the energy consumption, which is great because it's both economic. It's, I mean you make savings and at the same time you do good for the, for the planet. But reducing by 10% the, the consumption when you have a traffic on your network increasing by 30% every year, that's a huge challenge. We do it thanks to decommissioning the old. It works. We still have 4 mobile networks in France, 2G3G-4G5G2G and 3G will be decommissioned in the coming years. We have two fixed networks, copper, fiber. We are working. We've started decommissioning copper. That's going to take time because this is a huge network that it took 40 years to build. But we are on a journey to decommissioning by 20-30 and the one thing that we will have all of us to work on. Today, on average, a family in France has 10 devices. PC's, mobile phones, tablets we sell every year, not orange, but the industry in Europe every year. 150 million mobile phones are sold every year in Europe. Unless we extend the lifetime of mobile phones. There is no way the industry achieves neutrality. So this is of course a bigger question for Apple and Samsung, but we are working actively with them. We are one of the most active distributor of iPhones. I mean Apple and Samsung devices pushing them. But it's it raises question like, you know, we have for Android, I mean for Google, I mean, how long should you support the iOS software version for devices? And so there's a real question. Hardware, software management, if you want to extend the lifetime of mobile phones. So, so we have many, many initiatives. And maybe the one thing that is the most important for me, we are very committed to net neutrality. We are here in France, we talk about great things, AI, innovation. We also operate in countries where people are hardly surviving, hardly get food, they have no roads, no infrastructure. We are very committed to bringing mobile networks everywhere in Africa. And sometimes we bring mobile in, in villages where there's no electricity. So how do we continue to do that? How do we embed a I at best in Europe while also leaving no one behind in the in all the countries where we operate? And that's a bigger question, I think for our society. And a question about human you know, AI is not just a bad tech. It's of course a matter of, of people. And how does it impact you internally? Yeah, with your, with your employees, How do you train them to, to AI? So of course, like any large companies, I mean we're used to training, to training our employees. And this is something that's very important because I'm sure as I mean, if there are many startups in the room today, everyone is doing proof of concept. The big challenge is how do you scale use cases in a company? And the challenge is not gonna be technology, it's change management, training, changing processes, adoption. So that requires not just training, but also making sure people are not scared of losing their job. And, and we know for sure that there's not a single job that's not going to be changed with AI. Whether you are a lawyer, whether you're in finance, whether you are mean the technician, whether you are in marketing, in sales. So we all need to make sure that we get used to it. And obviously things stopped by the top. So we have to lead by example also as as leaders, but we also have to be prepared and it's not easy because we don't have all the answers. I cannot go and talk to our Workers Union and say nothing will happen, we will not change jobs, will not be impacted because that would be wrong. The reality is that we know that some jobs will disappear and that we know we will create other jobs. So we have to create, of course, a framework where people feel safe that as a company will do what's right to make sure that they get trained, that they then can evolve to other jobs. We already know you take cybersecurity, which is something that we know for sure, which by the way, is impacted by her on the good side and the bad side. But we know AI, he's just, I mean, increasing massively. The the, the ability for attackers to attack companies and to create frauds for consumers. So we are increasing jobs in cybersecurity and we are embedding AI to, to make sure we can do that. So there's many, many things we do. The one thing we've done and that's done by Orange Business is that we have a lot of training internally that we use for us. We've also opened to, you know, SMB's to our customers the ability to use our training because the biggest challenge of course large companies, we have the means we work. With the artifact, we work with many players in the house. We have Orange business as well, But you know, you're the CEO of an SMB, a small company, you have to deal with getting the order book right. You have to focus on hiring people. I mean, I is the last thing you need. So, but we have a number of training that are easy to consume and we've opened that as well for for our customers. And you have you been trained to AI? Did you do? Are you using Chegg PT daily or the one thing? They have a good city who's challenging us in meeting room. But the the best thing is your I mean, I've always been surprised when I use so in in in house as well. For employees, we have a tool. That we call Dino 2. And it's not the New 2DS. I know US, no, it's DINOTOD U2 and it's a enterprise GPT. But we've embedded the missed call as well because we want to for French language, it's much better. And so we, we and we have more than 30,000 employees who've used the tool and some of them are using it on a daily basis. So that's the best way to get used to it. But I must say that we've tested GPT for all on a few things. We're actually even as executive sometime you challenge your ability to bring value. You, you know, you ask a question that you could ask me here on stage and you ask the tool, what would you answer? So it's it's putting pressure on us as well to bring with some ideas are not just politically correct copy paste from what we've said elsewhere, etcetera. So it's also a good chat.