Peter Weckesser: Making an IMPACT with AI | EP22 | Schneider Electric

I believe the first thing is to acknowledge that there is a relatively good chance that AI will have a significant impact on almost any business, and then any company needs to ask the question: what is that impact and how can this be steered into the right direction in order to have the best benefit from that.

Thanks, Gosha. Thanks for having me.

So yeah, thanks for that very interesting starting question. And I can clearly say in all the conversations that I am having with customers, that I'm having with vendors and partners, AI is top of the agenda. And that actually has been the case at least for the last two years. Very honestly, it really started with the emergence of generative AI and large language models. But actually the conversation is broader than generative AI only. What we see today is that all software vendors are really creating an AI-enabled portfolio. And we see this really all over the place. And of course, all companies are asking themselves the question: how can they utilize AI in the best possible way? And we see this very broadly across all industries that we are serving at Schneider. Of course, we are deploying AI at scale at Schneider ourselves. And for us at Schneider, it's actually both. It's generative AI, which is really about creating new content from mostly unstructured data, but for us, it's also what we call analytical AI or what was called machine learning in the past, which is really creating insights into mostly structured data. And we have use cases where we can combine both. Yes, very much a boardroom discussion at partners, at suppliers, and at Schneider as well.

So maybe let me start answering that question with a bit of a bold statement. My personal belief is that AI is today where the internet was about 30 years ago. And if you think of the internet for a moment, the internet has pretty much changed every company, every business process. It has changed all of our personal lives, right? We couldn't easily order pizza anymore without the internet, nor could we find the way to a meeting that we drive to. So the internet is everywhere and it has significantly changed how companies operate. It has significantly changed our personal and private lives. I believe AI has the same potential. And it's really what we are seeing right now is only the starting point. We can already see use cases where AI significantly impacts the way how we operate. Maybe the single biggest use case today is in content generation in software engineering, which is not only adopted by Schneider but pretty much across any company that generates software, because the new GenAI capabilities are really capable of supporting software engineers in their daily work and significantly drive productivity. The numbers that I'm hearing range between 10 and maybe 40% productivity gain for software engineering. We are clearly seeing that productivity gain in Schneider. And we have decided to roll out the respective tools to support, and we have about 10,000 software engineers in the company. And we have rolled out the respective tools to support all of our software engineering community, and our goal is of course a faster time to market. So we want to release new products more quickly, and this is why this is a very important tool and platform to become better and serve our customer needs more quickly. Now we have many other examples where we cannot simply use off-the-shelf tools but where we need to train foundational technologies to really address bespoke cases that we have in Schneider. So also the future I see this will be a mix of more off-the-shelf technology but also where companies have to tailor the AI solutions to really create more differentiated value for them.

So very interesting topic, and very clearly the world is going into agentic AI. What does that mean? This means that it will become relatively easy for everybody to create an AI agent, which is an AI-based application that delivers a certain outcome, a certain value. It usually needs data, it actually always needs data to either create some new content or create some insights into that data. And more and more companies are providing either AI agents or even tool suites that it becomes relatively easy even for people who do not have programming skills to create an AI agent. Now, here I have a little bit of a view that this is something that needs to be governed because not everybody in every company has the skills to create new software applications which include AI. And when I mean the skills, I'm not meaning the technical skills because the barrier is lowered so that almost everybody can do that, but putting a software application into operation certainly requires the necessary quality assurance. It requires a certain governance. Are we really doing the right thing? So this is clearly something where I want to put the right governance around so that we deal with this in a very professional and in an ethical way at Schneider.

So I believe the first thing is to acknowledge that there is a relatively good chance that AI will have a significant impact on almost any business. And then any company needs to ask the question: what is that impact and how can this be steered into the right direction in order to have the best benefit from that? And there are multiple dimensions of questions to be addressed. What are the use cases where AI can play a major role? Where will AI potentially change the way of how we do business? Where can AI drive pretty significant productivity? What is the change management that is required in any company? And also what are the technical skills that you have to build or acquire in order to master that? Now my personal experience is that the technical problems are not the hardest problems to solve. Really, these are usually the relatively easier ones to solve. That's also our experience at Schneider, that we actually are in a very good shape already to deal with the technical challenges. We have created an AI organization which has around 300 people. We have built up the technical skills, but we have done way more than building the technical skills. And this is really where the rubber hits the road, because our ambition at Schneider is not to play with AI because it's cool technology, but our ambition is to drive significant business impact through AI for the company and really use it at scale. And in order to do this, we have decided to drive a use case and business case driven approach across the organization. Everybody in Schneider, that means every function that could be finance, that could be manufacturing, that could be HR, or every product organization that owns a product becomes a partner of our AI hub. And the way how we work is that these spokes really own the business case and the AI hub is responsible for the technology platforms and the delivery of the business case. Now the experience that we also have is that you need to put quite a bit of focus on the necessary change management, particularly if you want to deploy AI to optimize internal processes. It's very important that we put the right focus on driving the necessary process change, which includes first informing people, educating people, upskilling people that some of the work we do will be done in a different way in the future. Yes.

Yeah. Thanks, very good question. And I want to actually not only give you one example, I want to give you two examples which are a bit different. So the one example which has most likely the biggest impact on our productivity was our decision to introduce an AI-based technology that allows our software engineering community to get significantly more productive in software engineering. The way we write code, how we test code, how we deploy code. Here we see certainly double-digit productivity levels across our software engineering community. Now these are off-the-shelf tools. They are relatively easy to use. You don't need to have any specific AI know-how to deploy these technologies. So these technologies can be used relatively quickly by everybody. Of course you need to adapt your processes and fully embrace this. Our ambition at Schneider is to become much more efficient and faster how we deliver new products to the market, and many of our products have a pretty significant software content. So software engineering is a key value contribution to the value proposition of our products. This is one example. I want to give you a second one which actually required a bit of tailoring and bespoke engineering to make that happen. That is a use case where we are supporting our customer care organization, helping to answer questions from our customers more efficiently. So with thousands of customers, we have a lot of customer requests when customers come back to us with questions on pricing, on availability, on technical questions on the product. And these questions usually go to a customer care agent that gets back to the customer with an answer to that. This can be via many different channels: phone, email, chat technology. So we are using all channels for that. And of course with our very broad product portfolio, these customer care agents need to be trained. And they still need to have a knowledge base that they can easily search. Now we have created a chatbot which is based on generative AI which allows our customer care agents to find the answer to the specific question much more quickly and efficiently. So we have been able to reduce the time spent on answering a customer question by up to 50%. This is a huge efficiency savings and it's a huge customer value because we are able to get back to our customers much more quickly with the right answer that they are looking for. So it's serving both dimensions: customer satisfaction as well as productivity improvement for ourselves.

Yeah. Yeah. You're asking the question and giving the answer already, at least partially right. So for us, it's not the number of projects and it's certainly not playing with technology, but it's around measuring the impact. And there are some of the hard KPIs which are impact in terms of what are the productivities, efficiencies, cost savings that we can generate. We are also very consciously looking at how can AI drive more topline, so more customer success with our customers. So how do we make our portfolio more attractive? So these are two dimensions where we created a principle: before we embark on any AI project or use case, we want to be very clear about the metrics how we measure the success in this. And usually these fall into either of these two categories: productivity and efficiency or into increased topline. Some of the use cases are really actually driving both. So these are the more hard measures. Then I also believe, because we clearly have the ambition to be the most digital company in our industry, that we also need to measure the satisfaction of our people utilizing these new AI capabilities, because I want to make sure that we have the most motivated people, people that really embrace this at scale. So this is part of our KPI system that we want to deliver the best experience for our people, and with this it's important to measure the satisfaction.

Yeah, thanks for that question. So for about six months I have had the honor now to serve as the President of Digital Europe. Digital Europe is an organization that comprises of 130 corporate members plus about 70 trade organizations from all the member states of the European Union. And Digital Europe has really become a major voice when it comes to the digital transformation of Europe and the competitiveness of Europe. And this is already one of the key answers. The ambition of Digital Europe is really to help bring Europe back on a level where we are competitive with some other geographies. Take the US, take China, take India. And clearly numbers show that Europe has fallen behind in the last 10 years. We need to step it up in Europe, particularly with the recent political changes. It will be even more important that Europe really starts to stand on its own feet and gets into the control of its own future. And I believe in this whole context of Digital Europe, it's very important that we look at Europe as one marketplace, that we step up the investments into technology and into the deployment of technology. Just want to reference the Draghi report that very nicely points out the necessary investments that need to be done in Europe. And then of course it's called Digital Europe, right? So because we believe that digital is a huge driver of any development economically in the foreseeable future, and we need to also master and own these digital technologies in Europe. This means not only that we create some of these technologies, we also need to deploy them at scale.

Yeah, so I think all experts agree that AI needs a certain level of regulation and this is not unusual. There's a good number of industries that have regulations and that need regulations. Think pharmaceutical, aerospace, these are clearly industries that depend on really having a smart regulation in place. Now I personally believe that AI also needs regulation, but we need to make sure that we create the right and smart regulation to prevent unethical use of AI, but also enable the use of AI to make European industry more competitive in the future. And there is certainly an area that we need to discuss more in Europe: that our ambition is not to create as much regulation as possible. But the ambition for Europe in the future should be to create as much regulation as really necessary. But let's make sure that we create smart regulation that creates the right boundary conditions for the right use of technology, but also enables European industry to be competitive on a global scale.

So Europe has an AI regulation. And this AI regulation is a risk-based approach, right? Where for any AI application that you need to put in operation, you need to go through a certain level of risk assessment. I think this is an okay approach, but we need to handle it practically so that it doesn't overburden the usage of AI for applications. And just want to throw out an example: when it comes to optimizing the water usage in washing machines or the navigation of vacuum cleaner robots in homes, right? Then we need to make sure that this regulation doesn't overburden European industry and also users in Europe, that we cannot easily deploy these new technologies. And right now I see us at a pivotal point where we need to make sure that we do not create more regulations, but we make actually the ambition should be less regulation and make it smarter, right? And here we have some way to go in Europe.

Yeah, maybe a few very simple advice that I can share. First is I think probably a very conscious decision: is AI relevant for your business? And I believe almost everybody will come to the conclusion yes it is. And if you come to that conclusion, I recommend that you fully embrace this as an organization, not only in one department, but you need to embrace this AI transformation with your complete organization. And everybody in your organization will have a role in this AI transformation or at least be a user of AI in the future. So clearly recommendation is drive it from a use case driven approach. Which use cases have the most impact on your business? Then put the right teams on delivering these use cases. Do not try to create technologies that you can source from the market. Everything that you can buy from the market you should be using from the market. And yeah, fully embrace it and measure the impact and pivot if some of your use cases do not deliver the impact that you expected.

Thanks, Gosha. Really enjoyed the interview and maybe looking forward to doing another one in the future.