Implementing AI in the industrial sector: A conversation with Guido Kerkhoff, CEO of Klöckner SE
In this next episode of The Heidrick & Struggles Leadership Podcast, Heidrick & Struggles’ Peter Behncke speaks to Guido Kerkhoff, the CEO of Klöckner SE, a German producer-independent steel and metal distributor. The conversation revolves around AI: how it has changed the ways of working in the industrial sector, how it is being used, who is responsible for deciding how to deploy it, and how to manage the risks and opportunities it presents. Kerkhoff also discusses how to assess the performance of AI initiatives and how organizations can make sure they have the leaders they need to progress with AI and other emerging technologies and finally shares some advice for other leaders about how their organizations can implement AI tools.
Below is a full transcript of the episode, which has been edited for clarity.
Welcome to The Heidrick & Struggles Leadership Podcast. Heidrick is the premier global provider of senior-level executive search and leadership consulting services. Diversity and inclusion, leading through tumultuous times, and building thriving teams and organizations are among the core issues we talk with leaders about every day, including in our podcasts. Thank you for joining the conversation.
Peter Behncke: Hi, I'm Peter Behncke, partner in Heidrick & Struggles Frankfurt office and a member of the firm's global Industrial and Financial Officers practices. In today's podcast, I'm excited to speak to Guido Kerkhoff, CEO of Klöckner SE. After a stint at Deutsche Telekom, Guido stepped into the role of a member of the management board and CFO at ThyssenKrupp AG in 2011. He then served as the company's CEO in 2018 and 2019. In his current role as CEO at Klöckner SE, he is responsible for the coordination of the management board and functionally responsible for many corporate divisions.
Guido, welcome, and thank you for taking the time to speak with us today.
Today, I would love to talk to you about AI and to what degree this has implications for your work. So, to start with, how would you describe how the data and technology team at Klöckner worked before AI came to prominence?
Guido Kerkhoff: Well, first of all, thank you for having me here on this podcast. I'm really excited about the questions and to talk about our digitalization and AI experience. At Klöckner, we were in a very traditional industry—it's trading and it's metals-based. What is the key component of that industry? What makes it different to others? We have a large scale of products, so more than 200,000 different products and more than 100,000 different customers. So that's why data, how to collect it, how to get the right [products] out, and how to understand your whole value chain and the logistics around it have always been important. So, in the past, before AI and digitalization, we had already started to have big data pools so that we knew what we had, what was available, and what we can buy and what we can sell, because that's our role as a trader in this industry. So, data and technology have always been very important, but AI and digitalization have led us to a completely different level.
Peter Behncke: Did you have a culture that supported innovation and learning? For example, where are specific people focused on exploring new technologies? Which other teams at the company did you work most with?
Guido Kerkhoff: To really drive forward digitalization and go further with AI, you need people, first of all, that really do understand what capabilities you might have. Because people from the normal business, they do their regular jobs and very often lack the imagination of what is doable. So that's why we started very early to create special teams, especially in Berlin at our Klöckner facilities, where we have this special knowledge that can help us to understand what opportunities are out there, that really do understand how to build [on those ideas], and what to make. And we were early; it was around 2015 when we started. And what you need then within the organization is openness to innovation and learning so that you can distribute [knowledge to] and educate people. What we did very early from there on was to create a digital academy, where people indeed can get the learnings and the tools that they need online, and so train themselves to further explore it and not just have it offline where people have to meet. They can educate themselves. So we rolled out our digital academy very early and that was very helpful through the Covid times as well.
Peter Behncke: Is Klöckner using AI in many different areas of operation? In what areas is AI used the most?
Guido Kerkhoff: Where we found a very early use case for AI was our so-called Klöckner system. As I already mentioned, one of the large difficulties in our business is matching customer requirements (that can have a very wide range and are not standardized) to the broad range of what is available on the market, what the mills deliver, what different traders have. Nobody was standardized on that. So, if a customer placed an order, historically, a lot of the work was done just trying to understand what they really needed and what that meant from a mill side. So, to match the products and the availabilities to customer demand. And for that, we built (pretty early on) an AI tool called Klöckner Assistant, [to help with] these unstructured data that come in from the customer—because you can't force the customer to digitalize or to standardize anything—once the data came in via email or via PDF or whatever they use, we can read them out and the Klöckner system learns to match what customers tell us and what we have on our central database. That way, we can be much faster with our RFQs and replying to the RFQs and make an offer to the customer. And this system itself has an AI-based technology that learns while it's running. So, if the same customer asks for something the second time, the system already knows how it matched that last time, which is something it couldn't do the first time. So, that’s helped us a lot to be faster in front of the customer and to digitalize the process that wasn't digitalized in the whole industry before.
Peter Behncke: How did the company decide where and how to deploy AI? Who was involved and what were the considerations?
Guido Kerkhoff: In our distribution and trading business, the complexity is the main thing and that's why we saw an opportunity in moving ahead with all kinds of digital and AI-based tools to [add structure] and be faster—faster internally but faster for the customer as well.
But to grab these opportunities, we were very early. We started, as I said, in the mid-2010s. We built up our own team, our own department where we have the specialists that really do understand what is doable, and then we bring these people together with a specialist from the business to evaluate what you can get out of it. So, I think the starting base has to be really a dedicated group of people that develops opportunities and then will double-check with your specialists in the business on where you can improve and what use cases you might find.
Peter Behncke: Do you think AI is a bigger risk or opportunity in the medium term, and why?
Guido Kerkhoff: Clearly, I always try to be positive and see it from the opportunity side. All new technologies always have both sides, the good and the bad one, the risk and the opportunity. But, clearly, I think AI and all kinds of digital tools are the name of the game to further improve efficiency and to be closer to your customers’ needs, and therefore to progress with what you're doing. The whole economic life can only get better if you use these tools the right way, and that's what we're trying to do. And on the risk side, you always have to manage that. I mean, take a look at before, with social media and the internet—a lot more information was available but it can be the wrong [information] as well. So, as far as the way to manage risk—if you take a look, for example, at ChatGPT, that can only connect or use the data that it finds on the internet. If, within the World Wide Web, these data are distorted to some degree or biased, it means that the AI can only work with what it's fed. So, therefore, you have to control that, and you have to educate your people to still think and reflect before they use something coming out of [ChatGPT]. But I think it's doable. The same is true for data security. Yes, it is always a risk. But, you know, can you find ways around the risk by elaborating and working on the opportunities? I think it's doable. And that's always the challenge with new technology, so nothing new there. Maybe a bit more important now to really take care of the risks.
Peter Behncke: What specific skills or capabilities have been helpful to you and your team in working on AI with operating executives?
Guido Kerkhoff: Yes, I think [it’s important] to really have specialists that know what is doable, because there is such a broad range where you can use AI. For example, as I mentioned with our Klöckner Assistant, it was very specific on improving our processes. But if you take a look at all the opportunities that come out of ChatGPT—that can be from your strategy to your annual report to any kind of questions or analysis you do internally—it is such a broad range of things that are there, that, first of all, you need to really find the specialists and people that know what is doable, and then to connect them with your business people and the people who might have use cases. And it’s only when these people and the real specialists tell you what is doable that the imagination starts within your organization. And that's why you have to bring these teams together and then to find out on what use cases you can work and where it makes sense for you to explore further.
Peter Behncke: How are you assessing the performance of your AI initiatives and who's involved in those discussions and deciding what to do next?
Guido Kerkhoff: Yes, finally, it has to support the business. There are some places where it's easier and some places where it's a lot vaguer. If I take a look at this Assistant that I mentioned a couple of times now, it's easy to measure it with classical KPIs such as, what amount of time does it take to process an order? How fast can we deliver? How fast do we react? What about the percentage of RFQs where we can respond via Klöckner system and get the orders? Is it better than doing it the classical way? So, if you use tools to improve your classical and typical processes within your organization, it is easy; you can use classical KPIs. But, for example, if you want to assess whether your strategy gets better because your people use ChatGPT to analyze what is around in the world and, you know, what information you can process to get to better decisions, sometimes it gets a bit more difficult. The way I try to assess it is, all the tools we find and we use and we operate, do they really help us to develop a better strategy or be faster in the performance? Because they have to serve our business needs; they're not there for themselves. So people really need to present a use case and not play around. And especially if you take a look at certain of these tools, like ChatGPT, playing around makes a lot of fun as well but you need to bring it together to a use case.
Peter Behncke: What leadership skills or capabilities related to AI are you finding most difficult to find or develop, and now how are you making sure you have the leaders you need going forward with AI and other emerging technologies?
Guido Kerkhoff: Yes, that's indeed easier [said than] done. I think, overall, with the younger generations, a new understanding of technology comes in. So, it's very often a question of generations. The generations before us, well, they wouldn't be used to working with computers, but then computers came and people learned it at university and while they grew up. So, I think a lot of this AI stuff, when the younger generation joins, [the skills are] going to be there because they use it by doing their studies already at school and what have you, so they have a better understanding. But to really change your company and bring that even to employees that have already served for a longer time for the company, I think it is important to get this mix, to build these teams and to get these people that know, and to combine them with the stakeholders within the company that know your processes and know how you're operating. And then [we have to] form teams, bring people together, and try to develop the right use cases out of that. Because, very often, even older people who have been with a company longer are open to these new ways of doing things and you need to really bring together the right people.
Peter Behncke: Will employees with a different skill set be recruited in the future?
Guido Kerkhoff: I think that's largely going to depend on the industry you're in. If AI especially is something that can disrupt your business, then you might even be in a completely different business case and doing something completely different. In our case, I think it's not so much the case because we're in the physical distribution of physical material—steel, aluminum, and stainless steel—that will be needed in the future as well. So, the scope of our business will not change, just the way we do it. That's why I think, yes, we will need people who are natives to AI and all these digital tools, but it will come to some degree with younger and new generations as well. Therefore, to some degree, yes, [a different skill set will be needed], but we still need the experts of the physical world as well. So, the combination and, you know, what we try to bring together today will to some degree stay and younger and new employees will bring on the new capabilities to some degree.
Peter Behncke: What advice do you have for other leaders in working with executives around their organizations to implement AI?
Guido Kerkhoff: Move early and get experts from the outside. If you don't have them internally, get people who can really focus on these tools and technologies that can tell you what the space out there is, and then bring these people together in a second step with your internal experts who know exactly what you're doing and where use cases can be. Bring them together and show them what could be doable, and how you could improve. And to really make it happen later on, I think, one last thing is important: you need to make it a topic for the C-level because it will disrupt; it will change how you're doing business. And unless people have real support from the top to drive change and use these technologies for the best of the company, it will happen later. So, I think that's it: get experts, bring them together with the people, and support it from the high level, and then celebrate successes.
Peter Behncke: Guido, thank you very much for taking the time to speak with us today.
Guido Kerkhoff: Thank you.
Thanks for listening to The Heidrick & Struggles Leadership Podcast. To make sure you don’t miss more future-shaping ideas and conversations, please subscribe to our channel on the podcast app. And if you’re listening via LinkedIn, Twitter, or YouTube, why not share this with your connections? Until next time.
About the interviewer
Peter Behncke (pbehncke@heidrick.com) is a partner in the Heidrick & Struggles’ Frankfurt office and a member of the firm’s global Industrial and Financial Officers practices.