Skip to content
Featured image: Opportunities For AI In The Insurance Industry

Opportunities For AI In The Insurance Industry

Dr. Sebastian Kaiser, Head of Machine Learning

"AI is an exciting thing, but it has to be used at the right time and moment. I've seen people produce true magic with AI, but it boils down to that AI should be considered an enabler to something bigger."

In this episode, we welcome Dr. Sebastian Kaiser, Head of Machine Learning at ERGO Group AG - one of the largest insurance groups in Europe. His passion: Data and statistics. His mission: To relieve employees of repetitive tasks by using AI to enable a more customer-focused way of working and thus increase company growth.

By means of some concrete use cases, Sebastian explains to us how and where AI and Machine Learning are already being used at ERGO. We also learn about the bright minds that make up Sebastian's team, the 'Advanced Analytics Unit', and the central role behind the "AI Engagement Manager".

If you want to dive deeper into the use of AI in the insurance industry, but also if you want to know why Sebastian perceives AI and Machine Learning rather as a kind of "enabler for innovation" and where he thinks the market for AI is headed by 2030, this conversation might be worth listening!

Below you will find the full transcript for the episode.

Opportunities for AI in the insurance industry

Chris: Hi and welcome back to Innovation Rockstars. My name is Chris Mühlroth and in this episode, I am excited to welcome Sebastian Kaiser at ERGO. Being the head of machine learning, Sebastian and his team develop business cases from a large amount of data in order to create more freedom for employees. Now this clearly puts the focus on customer interaction. Repetitive and simple tasks can be automated and both customer loyalty and business growth are supported. So that sounds like an exciting spot to be in. Sebastian, thanks much for joining.

Sebastian: Hi Chris, thanks for having me.

Chris: All right, so let's kick things off with a 60 seconds introduction sprint, all about you, your career, and your current role at ERGO. And I do have a stopwatch here. So for the next 60 seconds, the stage is all yours. Let's go.

Sebastian: Hi, I'm Sebastian. I'm a data guy living with my family of four in Augsburg. I was fascinated by data my whole life. Starting as a statistician, I have a diploma in statistics, I have a Ph.D. in statistics, and I worked on unsupervised learning on big data - a very nice topic in the new development of AI and ML. After my PhD, I did consulting, and worked for the German Landesbanken as a joint venture where I developed rating models and brought them into production. Then I went to the data lab of Volkswagen, helping all the brands to use AI and machine learning. Then Munich Re caught me and I was five years at Munich Re. And now since eight months, I'm Head of machine learning at ERGO.

Chris: That's a great track record. All right, thank you. So next up, I have three sentence starters for you and I would like you to complete those sentences. So number one goes like this: “The difference between artificial intelligence and machine learning is…”

Sebastian: Well, there is not really a difference, right? So there is AI and machine learning is a part of it. It's not everything. There's a lot of other stuff in AI, but there is no AI without machine learning. But there is machine learning without AI. 

"There is no AI without machine learning. But there is machine learning without AI."

Chris: That's a great answer. Number two: “The most common myth about machine learning I encounter is…”

Sebastian: Well, I'm a little bit annoyed about the fact that people always say that machine learning makes stuff impersonal. So that the personal interaction is not there anymore. What we try to do at ERGO is to keep away the employees from repetitive tasks so that people can really concentrate on the customers and make it even more personal. I know that was more than one sentence, but that's what I wanted to say.

Chris: That's fine. All right. And I'm curious to hear your reply to number three: “What I have always wanted to say publicly is…”

Sebastian: Well, I always stay with machine learning and AI here. And AI is really powerful, but it has to be used right. At the moment, people always want to do magic with AI. But what they really should do is use AI as an enabler. To develop its full potential, it needs to be used right. And that's my opinion. We try to do that. Me, my team, and the whole unit. And hopefully, we do it.

Chris: Great. So that's a no-bullshit attitude. I like that. And you know, Sebastian, fun fact. On your LinkedIn profile, I found that you know the language Klingon on the level - and I quote - ‘elementary proficiency’. Now, if I understood it correctly, the Klingon language is a constructed language spoken by a fictional alien race, the Klingons, in the Star Trek universe. So, Sebastian, tell me, is this a skill that a few more of us should adopt?

Sebastian: Well, I hope so. So to say ‘Kapla’, which is a kind of greeting in Klingon, I'm a Star Trek Next Generation fan, and the Klingons play a big role there. That's how I developed that. I had an alarm clock waking me up with a Klingon sentence every morning. And so I was interested in that language. And since we are doing natural language processing now, it's always interesting to follow up with a constructed language like Klingon. There are also translation services that translate English to Klingon. The fact is, it's not based on machine learning because there's not enough data.

Chris: All right. So is there any tool available that actually can translate from Klingon to English in real-time? Does that exist?

Sebastian: Yes, it exists. It is not perfect because it doesn’t use machine learning as the other new translation services out there, but it is there.

Demystifying machine learning 

Chris: All right. Cool. Maybe I should give it a shot. So let's talk about, you know, technology first and then go, you know, gradually go deeper into machine learning and then also the insurance industry and then your mission and your work at ERGO. So, you know, Sebastian, these days it is pretty easy to throw around a few buzzwords when you want to get some attention specifically, you know, on LinkedIn and social media and stuff. So let me try this: Drones, Hyperloop, blockchain, quantum computing, nanomedicine, 3D bioprinting, artificial intelligence. And I could go on and on like this for quite some time. And artificial intelligence and machine learning are kind of joining the ranks. So there is a lot of noise around all those trending technologies, but my assumption would be that actually very few implemented them or deployed something successful with it. Why do you think this is the case?

Sebastian: Well, first of all, there's a difference between these buzzwords. Things like quantum computing and also AI and machine learning are there for a long, long time. There's a theory developed for at least 100 years now or something. So we now have the possibility to use it. Quantum computing aside, AI and machine learning are really adopted now. However, as I stated before, you need to do it right. It's not that you get a tool. You need the right circumstances. In order to have that, there are many exciting ideas around AI and machine learning. But if you just do that with AI, it's not working. You need data. You need digitization around that. Software, hardware, everything. And it's changing now. More and more companies see how this is going. But first, they concentrated only on the AI side. Now they see that they also need to concentrate on processes and everything. We try to do that. We have a very nice team here at ERGO, which is very nicely constructed. We really try to focus on AI as an enabler, which fits into the landscape that we already have.

Chris: And talking about that landscape, you know, one thing with AI specifically is that, you know, kind of everything is labeled as AI these days. So machine learning = AI, big data = AI, clustering = AI, role-based classification - well, you have guessed it - it’s AI. So, you know, based on your opinion and your experience, what qualifies something to actually be an AI?

Sebastian: Well, the definition was done before, right. It's not from me. I really like the famous British scientist Alan Turing, you might have heard of him; he has this very nice definition of AI, which tells you if a human interacts with something, and he does not know if it was a human or a computer when you ask him later - and it was a computer - that's what I would call an AI. I completely understand if a startup wants to label their stuff as AI and the European Union in their definition also lowers the level for an AI. So please let them call it AI. In a corporate, when we are talking about AI, we really try to do it. All this stuff that you heard about deep learning and learning of the machines and everything. This is what we do on the machine learning and the AI side.

Chris: Okay, that's clear. Thanks for describing that. So on the one hand, we got AI. And then, on the other hand, we got all sorts of discussions around corporate innovation at the same time. And you can, of course, argue that AI is a great innovation, but maybe not. So what is your take on the relationship nowadays between artificial intelligence on the one hand and innovation on the other?

Sebastian: Well, AI is an enabler. It can be seen as a technology or as a software kind of thing, but it is an enabler. And for sure, since it is new, it's kind of an enabler for innovation, but the two things are also completely separate. When we in insurance do innovation, that can be done completely without AI. We have a very nice new product on the dentist's side; when you have dent problems, you can even make the insurance or sign the insurance after your first treatment. There's no AI involved or anything. On the other side, we use AI to just improve our standard processes, e.g. in the back office of insurance, to root emails, and to read from documents automatically - there is no innovation involved in that. That was done before with standard technology, and now it is done with AI. But the nice examples are the ones where AI is meeting innovation, right? Because then it's the full power, and you really can change something in the world. What I really like in the insurance space on the innovation side, for example, is when you take a picture of claims. If you had a car accident, this is automatically processed to take away the stress from you. So whenever innovation meets AI, a lot of cool stuff is coming out. If you really bring it into operationalization and not only draw it as a mock somewhere.

Chris: Yeah, and just hang on the walls and say, well, we could do that if we would have that and so on. That's super interesting. And I guess we'll get to a couple of more use cases in a later stage of that episode. But before we talk about machine learning use cases at ERGO and also the insurance industry as a whole, let's play a quick game. It's called Either-Or. So, Sebastian, this is how it works. I will give you two options and you choose one and then spend one sentence each to briefly explain your choice. And I'm really curious to see what your answer will be to those questions. OK, number one: Do you see AI either as a friend or as a foe in the future?

"AI is a friend. We cannot really tell what's happening in the future, but if we train the people right, if we educate the people right, if the people know what they are doing, I do not see any problem with AI in the future."

Sebastian: Well, AI is a friend. And I have to spend two sentences on that. So what you see is that you already benefit from AI, right? On your mobile phone, everything. There are small applications that really help you, e.g. if you take photos it automatically finds what's in the photo. So at the moment, it helps you. We cannot really tell what's happening in the future, right? And AI can be abused. But if we train the people right, if we educate the people right, if the people know what they are doing, I do not see any problem with AI in the future.

Chris: That's right. And in fact, AI on a smartphone, for example, is already an AI in my pocket, right? So I'm using it all the time. OK, number two. I think you like both baseball and running if I'm not mistaken. So if you had to choose one sport for the rest of your life, what would you choose and why?

Sebastian: Baseball, for sure. I mean, running is part of baseball, right? So that's easy to say. And baseball includes a social element. Because other people play with you. And on top, there's a lot of statistics in baseball. And as you know, I really like statistics and numbers.

Chris: OK, that's a smarter answer. And finally, number three: Would you either get rid of your Ph.D. of your doctoral degree or stop working in the field of AI and machine learning?

Sebastian: Well, I think also an easy one. Get rid of the Ph.D. I do not even use this doctor's title in my name. I do not have it in my passport or anything. It was really nice to do research, and it played out really well for my career, but what I do now with AI and machine learning is so much fun, with nice colleagues, everything works out. So for sure, I would choose to get rid of this Ph.D.

Driving machine learning in the insurance industry

Chris: Yeah, you don't really need this on your doorbell nameplate or something, right? It's about the experience and the learnings you make during that time. Great, I knew this would be an easy shot for you. OK, now let's talk about machine learning at ERGO for a while and also the broader insurance industry as a whole. Now, obviously, as you said before, you're not a one-man show, right? You have a team around you. So let's start from there. How is the team positioned and what is the strategy of the team?

"We have a clear responsibility and an own budget. So what ERGO says is, look, we spend some money. We invest in that. You have to build something which brings then business value."

Sebastian: So that's something that convinced me to join ERGO - how the team is set up. It's an advanced analytics unit, as we call it, and it has different parts. There's a full data engineering team who do the ETL processes, providing the data and everything. We have a full data science team. We have AI engagement managers, which is really nice. These are people who engage between the AI or data science and the business unit as well as the IT. Sometimes a translation is necessary, right? And they are not an own team. They are with us in that case. Also, we have dedicated IT resources. This is quite nice. We have an own virtual private cloud solution, the AI factory. And dedicated resources just talking to us and developing that for us, which is very nice. So very strong connection to business and IT. And coming to the strategy, we have a clear responsibility and an own budget. So what ERGO says is, look, we spend some money. We invest in that. You have to build something which brings then business value. So the strategy is  that we bring some money with us when we're suggesting a new use case. But the use case in the end has to pay off. That's the idea. And everything around advanced analytics, not only AI and machine learning, also traditional analytics is included here.

Chris: That's super interesting. How many people is it today?

Sebastian: We are around 25 people, and growing.

Chris: Wow, that's cool. And you caught my attention with the AI engagement managers. That's an interesting role. Can you just spend some sentences on how they interact? Do they proactively look out for individuals in the organization? Do they connect or are they getting requests from the organization? How do they operate?

Sebastian: Well, it's really meant in a way that they do everything. As you know, data science teams, data engineering teams are built of math people. And sometimes there needs to be a translation between what the math people are doing and what the business people are doing. And this is exactly that role. So they spend time looking around for new use cases, but they also spend time helping to solve conflicts, to go to IT, to do all the planning. So the role is really kind of a generalist. They have to know what AI and machine learning are, but they do not need to program anything on that side. Also, they have to know what's up there in IT and also in insurance, but they do not have to specialize in any of those.

Chris: Got it. So they build bridges in some way, right?

Sebastian: Exactly.

Chris: Okay. And how much is the budget?

Sebastian: We are not talking money ;-)

Chris: But you do have a good budget, so it's not five euros or 500 or 5,000. You have at least a considerable budget to operate on, right?

Sebastian: Exactly. So we have this AI factory, which is running. And this is really a productive virtual private cloud. We are using Amazon. That's no secret. And we are paying that on our side in the team and also the resources we bring in. There’s a strong commitment of ERGO to go with AI and machine learning. Let’s frame it that way.

Chris: Yeah, it sounds like that. 25 people plus growing plus budget plus all the operation costs and stuff. So yeah, this sounds like a considerable investment. Okay. And talking about the team, how do you build such a team? How long does it take to find the right skill profiles? And specifically nowadays, how to hire, where to hire? What are some of the tips and tricks to build teams as you did?

Sebastian: So what ERGO, in my opinion, really did right is that mixture of the team. It's not that we only hire AI experts all around the world, placing them in Nuremberg and Düsseldorf. There's a good mixture in the team on the age side, on the experience side, also on where they are coming from. We have people who are with ERGO for more than 20 years. So we looked around at ERGO asking who is interested in these new topics, AI and machine learning, come on and join us. You get training there, you get educated. For sure, there must be some kind of interest in the topic, but that should be included. And I have not seen that in the other positions I had so far. There were a lot of external hires, and here there was tried to have at least half of the team from the inside. Very, very nice idea. Also good management, right? There is not an AI guru being the top manager of our unit. There is somebody who is for more than 22 years at ERGO, and he is managing the team as if we are a real ERGO unit. Nothing special, a real unit, just doing something new. This is very good. And also for sure, we also got people from consulting, also innovative people on that side, to have a good mixture. And I think this is the important thing. Another point, always you need to give the AI guys as much freedom as they need in order to have them there. For sure, you need a certain degree of freedom, but it should fit to the organization. And insurance is conservative, a little bit, right? So it also has to fit to the conservative side and to the big company side. And that's what I think works in our case.

Chris: Yeah, OK. So you got a lot of that covered. You get breadth, you get depth in skills and capabilities, and you get some managers who know the organization from the inside out. Know about culture, know how things are done, how decisions are being made, so it's in that mix of building an effective team inside, as a traditional, but of course, also a very ambitious organization.OK, so now specifically to you, I got to know you as someone who quickly cuts through the noise, right? So as Head of machine learning, it is about developing effective and also real-world business cases and real-world use cases. This needs to be applied. It's an enabler. Things that you already said before. So how do you and your team approach all of this? How do you operate? For example, where do you get your ideas from? How do you decide whether an idea shall be pursued or not? How do you work?

Sebastian: So the advantage at the moment we have is that we really try to work output-driven. We are looking for where we can make an impact, but we do not need a full firework or anything. At the moment, it's OK for us to just provide impact. And that means we stay very close to information technology. Very close to the business units and say, where are the small pain points you have where automation, optimization, everything that makes ML really powerful help on that side. What sure needs to be followed is always to be very near to new technology developments. But you cannot bring the newest technology into a running system. There always has to be an adoption side and we are not taking the very newest AI method and just placing it into insurance. We look at what really fits in there, what can make an input, and put it in a development process. For sure, insurance is all about the customer, right? Because you need customers, you have their money, they trust in you and everything. So we try to make a little bit more money with the AI that we have as it costs. That's where we start. And it works out well at that point. Very good, I have to say at that point in time.

Chris: OK, that's great to hear. So do you really write a business case, like, this is maybe the savings, this is maybe the cost advantage, or this is maybe even an assumption of additional revenue? Do you kind of go at that depth or are you operating in a different mode?

Sebastian: Exactly. That's what I meant when I say AI should be also used in how the company is working. We do all business cases. That means AI is no exception. I know that in a startup or in these new labs - I worked in labs before at Volkswagen and also at Munich-Re - there you do not look at the business cases, you look at the full potential. In ERGO, we for sure also have some kinds of things in mind that we do in five to ten years. But for the real thing where we are investing, we look into business cases and do it as a normal project. As we have done in the financial industry when I was there, a normal business case is that AI is just an enabler to get hopefully more money or more potential out of that.

Chris: Brilliant. And are you doing this all, you know, exclusively in-house or are you also open to collaborating externally as well?

Sebastian: Well, there's so much stuff to do in-house at the moment that we completely focus on in-house at that point in time, for sure. Always when you have something successful, a good manager has to ask, could we not externalize that? 25 people sounds like many, but it's not enough to even cover what's already inside. For sure, the focus is completely on the inside at the moment, but if there's something on the outside, I will not get away from that. Let's see what the future brings.

Chris: OK, fine. Now, you know, since AI still is considered to be maybe an emerging technology that's already at a certain point of adoption, for sure, but what do you need to consider when introducing AI to an organization? Do you consider this to be a change process? I guess there is a lot of communication involved. So, you know, how do you address maybe concerns or even fears like, oh, they're taking away my job or they want to replace me totally? What's your experience at ERGO with that?

Sebastian: Well, as you said, what you need to take away is the fear. I always wonder how people still fear stuff, although it's all around them. As we said, you already have it in your pocket. So why should you fear something that you put into your pocket? But this is clearly what you have to do when introducing that to the organization. And it does not help if you have a lab structure and just doing AI with some weird guys in hoodies sitting somewhere. You really have to go to the IT, to the business units and tell them, look, we are working like you. We are just using a different tool. That's what we are doing. And for sure you need a change, right? It's not that you do not need to change or the culture doesn’t need to change. But most of this change is already done because of other things. There's great innovation done on all this. There is digitization. You need digital processes. AI only comes on top. Sometimes people say, well, we need AI first. No, we need digitization first. And this is the cultural change. And for sure, this needs to be started. We have a very good Chief Digital Officer who came from T-Mobile to us five or six years ago. He started with digital and not with AI, which is really good for us now. And for sure, on the AI side, you need to do all this training. You need to train the people, you need to educate the people, you need to make them understand that AI is something everybody can understand because all of us did our own learning, right? We have intelligence. But this time, it's just not the guy sitting next to you. It's artificial intelligence.

Chris: So if I understand you correctly, you position this as an established practice. Is that fair to say

Sebastian: That's a very nice summary, exactly. We say it's a tool. It's already there and it works now. Sometimes we still have to prove it. Although there are so many proofs out there that there is a value in these new tools, AI and machine learning, you still have to prove it on the use case. And if we have to do so, we try to do so and convince the people.

Chris: That's brilliant. That's a great learning. Can we maybe get to some examples? How can AI and machine learning support, for example, customer loyalty and business growth? Can you talk about one, two, three examples?

Sebastian: Yes, for sure. But first of all, as said before, AI alone is not enough, right? It's not that you put AI on something and then you get more customer loyalty or more business growth. It always has to come with a digitized process or a new tool, an app, for example. The AI is inside that app or helps the digital process. One and the best example at the moment where it really works is the response time, right? So when a customer talks to an insurance, normally he writes a letter or anything, then waits several days until someone gets back to him. And this is something where AI can really help. You can route the email way faster when just scanning it and saying, this is the right person to answer that email. Letting pre-formulate the reply email and the human just double-checks and sends it out, makes everything much faster. Another thing is to make the communication outside the office hours. People do not have to wait to call the insurance. They can call you anytime. And there are chatbots who can take over normal smaller jobs. This is where AI is also involved. So AI makes customer contact faster. This is the main use case, I would say, at the moment. And also for sure, it’s cost effective and more individualized. People always think machine learning makes everybody the same. In our case, not. We can really look now into every single differentiation between people.

Chris: OK. And what would you say, do you have more internal or more custom related use cases? What's the distribution?

Sebastian: Well, the thing is that in an insurance company, you do not have so much internal stuff which is not related to the customer in the insurance value chain. When we are doing internal processing, it's always with the customer. So in marketing, you market the product. In sales, you sell the product to the customer. In underwriting, you make the price for the customer. In claims, you make the process for the customer. That's why I have to say mostly internal, right?

Chris: Yeah, that's great. And when we turn to the future and maybe also look at the insurance industry as a whole, but of course also at ERGO, how do you imagine the future is going to look like? For which business cases could AI and machine learning potentially be used later on? And what maybe does it take to make that possible?

"What we need in the future is more digitization. Once we have more digital processes, then there is really no border for machine learning and AI in an insurance company."

Sebastian: Basically, at the moment, we do optimization, right? So we optimize the processes, we get closer to the customer, we have additional features, chatbots, apps where things are automatically transferred. This is cool. What we need in the future is more digitization. Once we have more digital processes, then there is really no border for machine learning and AI in an insurance company. There is so much stuff you can do together with robotics and everything. On the marketing side, everybody heard about AI in marketing, right? With Cambridge Analytics, we do not want to talk about that, but there's a lot of stuff going on there on the marketing side. But also on the underwriting side, making more easy judgment of the risk, making it fairer. At the moment, everybody is afraid and like “Oh, the algorithm may be biased”, but who tells you that not the people who are building the algorithms at the moment are biased, right? We try to avoid that as much as possible, but you cannot really put into that. And AI can, because you tell him, please try to get that without bias. And so they will be much fairer. If you have a claim, look at this flood catastrophe in Germany when you call your insurance and nobody can answer the phone because there are not that many people working in the insurance and there’s too much claim. That could be solved in the future with AI and machine learning. You get on time answers and everything, so that this is really good. But getting to that, we first need to do all this digitalization, homework, and optimize the machine learning and AI processes. For sure, all of this stuff has to come in line with GDPR processes and everything. And to be fair, I really like that we have such a strong GDPR rule because then I know exactly what I'm allowed to do. That helps us, right? And this is quite nice.

Chris: Yeah, so it gives guidance and it gives some framework to operate in, right? So that doesn't necessarily have to be a bad thing. Totally agree. Well, that's super, super helpful. And thanks for guiding us through the noise of machine learning and artificial intelligence and for giving us some real life practical use cases. Now, let's talk about the entire market. How do you think the market for artificial intelligence itself will develop by, let's say, 2030?

Sebastian: Well, I hope that there will be a consolidation of what's happening now. So until 2030, there will be a lot of startups coming up, new companies. Our companies will develop tools on AI, as we just discussed before. If we solve something good, for sure top managers will say “Can we sell that outside?”, right? So the market is so big, and it's really hard to keep up with new technologies, new services, new tools, all the hyperscalers, like Microsoft, Amazon, they all offer new AI solutions. So I hope for a consolidation of the market, and I'm quite sure that it will happen. There will be more proofs on what is really working and what is not working. So you can hopefully easier say this is working and this is working for my problem. And this is good. Another thing is we do not educate enough people on the AI side. So this war of talent, which we have in nearly every business at the moment will become really, really hard because the universities are not completely ready for that. And there's so much stuff around data, not only AI, but what you need for AI: data engineering, data ingestion, all the AI engagement managers, right? Who have to be trained and educated. And that will be our biggest problem in 2030 on the AI side that you do not have enough people who know what to do.

Chris: Yeah. And the war of talent already begun, right? And if you just project this into maybe, I don't know, five or 10 years in the future, well, let's see where you can get other people from. That's going to be an interesting one. And what is next for machine learning at ERGO? So what are the next high priority topics on your radar?

Sebastian: Well, you know that insurance always has taken into account optimization and also automatization. So for sure, if you send something to the insurance, it was already an automated process 15 years ago. So that means you even have to be better than the systems who have been optimized now for 15 years. So there's a lot of things to do and we are very busy with doing that. And since we talked about the war of talents before, we need to educate our people. We need to start the education as early as possible. There are a lot of math people working in insurance because of the nature of things. They all are capable of doing AI programming in the future. So we start now, educate them, train them. And you need to keep up with the developments, right? Market disruptions for sure will also hit some of the older companies, like insurances, which are there for 150, 200 years. So you have to keep up with the new developments, data, software, hardware. That are the highest priorities for us at the moment.

Chris: So a lot of things to do. All right. So we looked into the future. Now, let's look back just for a second and let's look back on your professional career. Sebastian, what would you say was your greatest Innovation Rockstar moment so far?

Sebastian: Well, on the innovation side, I have to go back to my second to last employer, the Munich Re. So the mother company of ERGO, where I had the opportunity to work with a very innovative team doing Internet of Things. And what they were able to do is a paper part product. I don't know if you heard about that, but together with Relayr, which is an IoT company of Munich Re, and also TRUMPF, the laser cutting company, and HEIDELBERG, the printing company, we were able to develop a product where industry is only paying for the output and not for the machine anymore. This was done using data, digitization, financial knowledge, insurance knowledge, and for sure some AI and machine learning. And this is really an innovation. And there I really felt like someone who can innovate something.

Chris: And really gets an impact and output out of it, right? So that sounds like a true rockstar moment. Thanks a lot. And with this rockstar moment, we wrap up this episode. Sebastian, it was a pleasure to listen to you. Thanks again for joining. I really enjoyed this.

Sebastian: Thank you. It was also a pleasure for me.

Chris: All right. And to everybody listening or watching, if you like the show, then leave us a rating or a review and share the podcast with friends, with colleagues, with whomever. If you want to get in touch, simply shoot us a message at Now, that's it. Thanks for your time. See you in the next episode. Take care and bye-bye.

About the authors

Dr. Christian Mühlroth is the host of the Innovation Rockstars podcast and CEO of ITONICS. Dr. Sebastian Kaiser is Head of Machine Learning at ERGO Group AG.

The Innovation Rockstars podcast is a production of ITONICS, provider of the world’s leading Operating System for Innovation. Do you also have an inspiring story to tell about innovation, foresight, strategy or growth? Then shoot us a note!