Innovation Dividend 2.0 - How AI Is Redefining German Family Businesses

Speaker A: Hello and welcome to a new episode of uh, our podcast Beyond Core. The podcast for business leaders who dare to think beyond today's success and build a future. As you know, in every episode we dive deep into strategies and ventures of established companies and how they build new businesses based on AI and entrepreneurial execution. Our guests are industry leaders and pioneers who are shaping what's next. And today we will explore how AI is redefining German family businesses, the so called Mittelstand and our hidden champions. And our guest today is a special guest. It's Henrique, uh, founder, uh, from the corporate venture builder bridge maker and one of the biggest venture builders here in Germany. So we are kind of competitors but I think we are bringing different angles together here and it's always a very fruitful discussion for our uh, audience. And also important to know is that Henrike is co author of the book the Secrets of German Family Businesses and have spent years researching and advising organizations on topics like venture building. Hello. Hello Henrike.

Speaker B: Hi and welcome and thank you uh, for having me.

Speaker A: Um, before we go into the topic please can you share kind of a short introduction to our audience, who you are, what is Bridgemaker and what are you doing.

Speaker B: I think the basis when it all starts started with my mission with Bridgemaker is I was in succession planning and I was mediating some generation uh topics in family owned businesses and then also uh got some insights in the topic how we can make them future proof, what's their new innovation model. And then I founded my own company, um, and then I went into consultancy and then 10 years ago in Berlin I started Bridgemaker and as you just said we started as a venture builder and then I think with the mission to make these companies family owned businesses, German Mittelstand, future proof and help them with some innovation tools, uh, from venture building and today we are also advising since innovation is coming more closely to the core. Uh lately we are also advising on AI technologies, AI strategies, but then also implementing and building real AI products and also value creation in some private equity companies.

Speaker A: Very good. So it's a broad field and with 10 years so you're, your company is twice as old as ours. So I assume you have a lot of good stories to talk about today. Um, but first of all I mentioned also you are co author of the book the Secrets of German Family Businesses. Maybe you can share a little bit what was your contribution to the book and what was your kind of chapter about.

Speaker B: So I'm like um, I'm like on several advisory boards and family owned businesses and Also replacing some advisory board members. And that was also my role um, in that book like really bringing together. Since it's like the one part is there are 50 profiles or portraits of um, family owned business book really coming from their mission statement, um, their strategies, how they transform, how old are they. Like really this claim from tradition to innovation. And then we also have different essays um, in this book and my part was uh, to write an essay how AI is redefining family owned businesses. So how can family businesses apply AI in different use cases from automation. But also I think that's the beauty then ah, coming to new business model, ah, helping with A.I. technologies.

Speaker A: Um, but there's also secrets around German family businesses. So what makes them the hidden champions of our industries and why they are successful. And sometimes they talk about there's a long term orientation they have. They are kind of founder driven and with this long term orientation they can also long term can have maybe long term investments um, maybe have a bit more patience in building up ventures. So what are your takeaways from, from those uh, from this book as well from the other chapters?

Speaker B: Yeah, I think what you, what you just mentioned and I think that's the basis number one it's this long term orientation, right. It's based on they, they have sometimes like a strategic goal in mind which lasts longer than maybe in a, in a, in other companies which are like um, capital markets or so when you have a management and this is changing but in family businesses you have a long term orientation. I think this is basically number one. What is also secret is they have really good customer relationships which last for years. So you can also rely on data for example which is there. But on the other hand side it's also like a very personal relationship which they have in this customer relationship. And then of course it's on the one national uh, excellence. But um, of course especially in Germany we have like this engineering excellence which lasts for years. Um, and I think this is just like two or three samples what make them special.

Speaker A: Yeah, um, yeah, very true. So I also see this in our relationship. This engineering excellence, um, planning in detail, construction in detail, everything works. High quality standards, etc. Now we talk about AI and the impact of AI and it's about 1000 days ago, uh, since ChatGPT moment arrived in the world, uh, everybody's now using a kind of a LLM, um, making part of the, of their business. But they also need to think about new business models. So how do you see those strengths of the German Mittelstand of the Hidden champions like long term orientation, customer relationship, blah blah blah. How does it relate also into the AI era?

Speaker B: I think when we, when we come back to these three advantages and then I make an overall conclusion it's like if you say long term orientation but I think we have to take an eye on that we are faster in decision making also because we have like uh, other data points to make faster decisions due to AI. I think this is really important that we remind ourselves to be quicker in that. And I think especially family businesses since they are often controlled only by the family, it's able to do this faster decision making and also with the customer relationships. For example we have data driven perception. So um, we can like in yeah make make products and services more individual. So this is also something I think we have to keep an eye on or so engineering excellence. It's also we make products for example smarter. So so and I think in general when it comes to AI the question is how fast we learn and how fast can we adapt in this era of AI now.

Speaker A: So this also needs to then translate into the products and services. It's not just about automation of internal processes. So like yes we use AI to make quicker decisions but also as you mentioned we want to build maybe more personalized product or individualized products and services and therefore we also need to adopt our and our thinking about the product itself.

Speaker B: Yeah, I think the automation is just a necessity so it's a duty we all have to do. I come to that in a minute. But I think the beauty or the advantage is beyond core. So and that's why I also liked your, the name of your podcast and also your thinking about that it's beyond core because this makes the differentiation and this also defines new USP of the products and, and the company. So for example if we everybody has to start I think at automation to increase the efficiency to be uh. Yeah to be faster and we can also talk about some examples in a minute if you want. But then it's really beyond call based on new data insights. For example how can we use like for example infrastructure from the big giants like LLM systems or something and then enrich them or increase them with our or with the proprietary data which every mittelstands company has.

Speaker A: Hi. Impact of the Zapongja hut on the center yet. Uh uh, Um, Um. Um. Yeah and examples are always good because it makes it more tangible for the audience to understand what we mean by that. So yeah we have an example attempt

Speaker B: I think um, beyond core or in the core this is Just the first framework. And then when we come to maybe one intersection of framework and then I come to an example, we are always talking about innovation horizons. And um, when we started Innovation Horizon 1, this is close to the core. And I already mentioned that some of the innovation initiatives go closer to the core at the moment. And then you have a second innovation horizon, um, which is a mixture and then the third, it's really like innovative new business model in a way. And I think the automation, it's like starting with Innovation Horizon 1 and for example, um, really think about what you can automate. So and for example when I look at our own efforts, which we did there, I was thinking like three to four years ago, how can we automize everything which could be automized. So really that was also my mission that I said like and that was the goal. Think about every process we do and then automize it. And same as you just mentioned, we were in the incubation field. We are brainstorming ideas, we are conducting a lot of trend analysis and all of that. So starting that and then validating business model. And even especially these first phases where you have a lot of data, where you have repetitive tasks, this, this ideal to automize. And then it was that we founded our own AI like database for innovative ideas called tao. So this was really like an automation process. And this was for me personally kind of strategic move, um, because we cannibalized ourselves and we cannibalized our work at the beginning. Um, but I knew right now or even a year ago AI would do this anyway. So I wanted to be on the forefront of our innovation services. And really that's why I took the decision, strategically decision back then to automize that when I do. In conclusion right now I think we built too much ourselves. But maybe this is also we can talk uh, later about where to buy when built and then also when only adapt. But this was something in our own field, what we recognized. And then for example, um, automation of procurement disease, for example, um, when we think about how much time it's spent when there is a question from a customer, like an incoming question and inbound, and then the database is running, calculating prices, sending out a proposal, something like that. And this can really be automized in different workflows with AI agents like back to back, um, and really do it in kind of a system and then really think about this free human capacity, what you have there, what you can do, what you can do with that.

Speaker A: Um, yeah, there are multiple very good points you mentioned here. So it's a little bit like ambidextrous approach here. Right? So you need to apply AI in processes. I mean to drive efficiency, to drive automation but also to learn. I mean this is a safe space. When you apply it in your own processes then you can also a little bit like play around you prototype things. You can fail but you can also get familiar with the technology and the capabilities. But then the other part is also as you said beyond core thinking long term how is the impact to our business model and what else can be um, or what other business models we should build uh, based on AI? It's maybe also a bit like a succession, succession strategy. Right. So I mean you need to think about it, the future and what the future holds. So, so you need to feel the urgency to build something for the next generation or for the next months, for the next weeks, for the whatever how fast AI is. So in, in your conversations with the middle stand with the 50 companies, how do you sense how big the urgency is to think beyond building new AI based services?

Speaker B: I think it's pretty urgent in general to think about because what I also realized and I just mentioned in our own. So we build it up from our own. But the development of this big tech giants for example this is a commodity, right? So the infrastructure which you can use um to build then your own model on it, this is a commodity. And um, I think we are just on the starting point of all this technology. So it will be maybe faster, it will be smarter doing that. So I really have to think about what is commodity, what can other people do better and but what is really my USP in that and how can I apply this for example for bread products. So for example in another company we are working with we implemented an NI driven market radar. So it's a consumer based company and we are looking about for example nutrition trends um, and then based on that and also on the, on the customer inside what they are looking for example what they're scrolling, what they are looking for, content which blocks their readings or what they're interested in. And then we use these data to yeah innovate new products with it, new product combinations for example and then send it out. Uh, or yeah develop this and have this like yeah ah, have the markets insights like in real time.

Speaker A: Yeah, that's a good example. Um, I can also add one example. We worked together with a bigger retail company and they had an internal consultancy services service um or an external consulting service or external partners could hire the consultants from the retail company. To help them to improve margin and growth and stuff like that. And um, then they thought about this is hard to scale. I mean consultants, it's kind of a linear business so you need to hire more consultants and then you can offer more to the, to the world out there to your customers. So it's hard to scale. So why not trading kind of virtual consultants um based on our expertise from the real consultants or the, from the human consultants. So they are not replaced but they are now kind of duplicated and you can duplicate them 10,000 times. And now you have the ability, the capability to um, offer this consulting service to the whole market at once to a smaller price. But for you as a company it's hugely m scalable. And your previous consultants, the human consultants, they are not kind of continuing their job as a consultant but they are kind of supervising their new AI agent colleagues and train them and supervise them and maybe supervising also the projects a little bit. So they all get, also getting a new role. So it's really thinking about turning your assets into maybe uh, services and products

Speaker B: and maybe with this uh, to add to that um, since we both like ah, validating new business models, right? Innovating but then also validating. It's this I think what's really special about this testing things. It's this virtual Persona. Not only virtual consultant, what you just said, but it starts earlier virtual Persona and what I mean with that. It's like a uh, testing uh scenario. So you can conduct a lot of data combinations and a lot of new product combinations. And then with an AI dialogue for example with a chatbot you can simulate a lot of different product and service combinations and then you get really in real time also the feedback. So it's much more precise. And uh, especially when it comes to our uh, history we called the clients, we are in the field, we always tested the customer perspective on this, sometimes even not existing products. And this is now like in real time in a much shorter and much cheaper combination possible due to this like virtual Personas testing ground.

Speaker A: And it's also uh, absolutely required with AI and the new capabilities you can also wipe code a lot of products and services much much faster. So I say now everybody has the, to build shitty services and shitty products. So it's not just you can build it faster but now you can also increase the amount of bad ideas and bad products. So the, the importance of validating them maybe not with humans only but also in a way that you use synthetic users, synthetic customers etc. But you the point you need to Validate becomes even more important because yeah, you don't want to produce more and more of, of bad products and services. You want to be the real ones faster. So. Absolutely. Good point by you.

Speaker B: Um, also because the customer expectations are increasing. Right? Because it's what you said, like uh, building a basic product, but it can do it much easier. So the, yeah, the differentiation and really the diversification of the product and service portfolio of the companies, it's becoming much more important.

Speaker A: Exactly. Um, there are also a lot of well known services like predictive maintenance, etc. It's kind of common standard right now in the industries. But since 10 years, I would say since the whole topic of IoT came up, um, and everybody understood, okay, this is also a huge topic and everybody's talking about predictive maintenance. But do you see also other options, other services beyond predictive maintenance? Because this is very obvious and um, um, it's common standard already. But what, what else uh, could, could, could imit work on?

Speaker B: I think um, predictive maintenance is not enough anymore because what you just said, right. I can uh, I can maintain better and then I can even before an accident happened, for example, I can uh, intervene uh, in a way. So I'm coming from this like pull to a push. But I think that's not even enough because that's, this is just the knowledge that something has happened. I think the automation there has to go even beyond that, that it' prediction. Uh, but it's also happening in action automatically. So for example Siemens is using AI directly on a M machine with edge computing for example which goes directly in the Siemens coding app. And then um, the action. So if there is this intervention, the action is happening immediately like automatically. Something is like yeah, ah, due to AI, uh, it has to become a better, better outcome. So I think uh, the maintenance and the prediction is one thing but it has to be topped up I would say uh, with immediately actions. Uh, which are yeah like doing or for example other like very basic. But when it comes to logistics, fleet management, for example, um, if you have the real time data, for example where queue is or something or where there is a blocked road or something that is. It's in real time. It's a rerouting of these trucks for example. And it's like you don't have the human in the loop. It's automatically like activating a different thing. So I think like in. To sum it up like in predictive maintenance, really think about what you just said. We have it for years but really think about how can I Make it even better and how can I even improve it?

Speaker A: Yeah and this is also a huge topic right now when you look into the Silicon Valley and wherever people are working on AI a huge topic of self evolving systems. So it's not just like uh adapting a bit but in applying what I learned but also changing the system based on what I learned. So in the history or uh, uh. We went back to our IT department said I need this feature and then they started to work on a new feature. And then months later you have this feature but now you build systems which are kind of learning from the usage of the user or maybe other data points and then prove the whole usability, the whole features by themselves. So uh, you don't need to go back to the IT department. Um, um, um. And this whole self evolving topic is this really fascinating at the moment. And I mean the Mittelstand they have a lot of knowledge, they have a lot of expertise and a lot of data points. I mean they are kind of uh. Well equipped for this kind of new thinking. Right?

Speaker B: Yeah, I would, I would also say that. But maybe two things to that. When it. When it comes to like um. A business model. It could also be uh. A thing to think about how can I monetize my data? So what data do I have and can I maybe use it? Of course without names like anonymous sell this data for example. This is, this is one thing um. Which we look in several projects at. Um. How that. But what you just mentioned it. I think it's also like a risk in the, in the data structure because we can do a lot with new uh combinations and AI and we can either use existing database and don't train them or we can train them. But uh. The thing is that every company especially company have uh. To look at the quality of data which is there because in shit out I would say uh. First I really. And that's why what we often do we really have to look at the data. How is the existing data there? So for example we have one. One of our standardized products is commercial OS or commercial operating system where we um. Like automize commercial tasks, sales tasks, marketing tasks, content tasks, all of that. But it always starts with a either CRM or with the ERP system of the company. And then we apply like existing tools around that. But we always have to start with the quality of the core because if this is not sufficient of course everything which comes later than applying existing tools, maybe produce some content, maybe have automated mail campaigns but also like conducting proposals with its existing pricing database but it's always like, yeah, we, you have to check your data first.

Speaker A: Yeah, this is the homework you have to do. Right. So to prepare it and companies which have done this in the last 10 years properly, they are in a better shape right now, in a better situation compared to the other ones. We also said that the Mittelstand, they have um, huge expertise. Uh, they care about their employees and the employees they love to work for the middle stand for their companies because they are owner driven and long term orientation and points like this. And so the employees, they have a lot of knowledge, they are experts in their field like German engineering and quality and stuff like that. How does it relate to AI? So how could maybe the Mittelstand leverage the expertise they have in the people, in the humans and leverage it together with AI and so not replacing the humans saying oh, thank you very much for the last 10, 20 years, now we continue without you, now we have AI. It's more about how could a really good combination look like. So building on the strengths you have and um. Yeah, um, leveraging AI.

Speaker B: Mhm. I think one of the key points there is that you have access to this knowledge, right. That you can conduct it, that you take it out, for example, out of this like humans. So for example, when somebody is not working for the company anymore or is retired or even just is sick somehow you can like build better systems to get this knowledge or make this knowledge accessible. Right. This is, this is one point. And the other point is also um. And this is really critical when you want to come from pilot projects with AI, some prototypes, really into standardized products, but even more into uh, the integration and the rollout and the transformation. And there the key success factor is that you educate your employees in a way that they are really using what you have with AI in your company. So we try really. So for example, within our revenue model we always had something which is related to the success uh, of the outcome, what we do. So for example, when it comes to the employees, we have like a milestone, uh, or a KPI, which is really like the adoption rate of every employee. And as I mentioned, it's not only the, you can say like younger, older, more experienced, not. It's really like an effort which you have to do like in the, in the whole company. Right. And this is really critical uh, if you want to roll out an NI initiative like for the whole company.

Speaker A: I, I also see it in a way that um, AI, so when you look at the big LLMs, they are trained on kind of book knowledge. So what kind of knowledge is available in the Internet. So they consumed and internalized the whole Internet and they are now um, trained on that, on all the books out there etc, so they are on a level like a uh, student or whatever when they finish university. But what they don't have is this, this very special knowledge about special tools about special engineering or whatever. This is not really published in the Internet. So the LLM maybe knows a little bit about that, maybe can think about it a little bit but cannot fully understand things in detail like the experts do. But now you can combine it. So you have the expert knowledge and now you have also the capabilities of the LLM to speak natural language and go through unstructured data and find commonalities or whatever else. Now you can combine it. So this leads to an advantage. But you also need to think about how can I protect this a little bit so keep it. So I don't want to share everything with the LLM and then tomorrow everybody has it. It's more like my expertise. How can I also kind of protect it a little bit and then combine it with the LLM power and then build some, some something on top so it gives them also the opportunity to build an advantage compared to LLM stuff or AI.

Speaker B: Totally that, that's what I meant at the beginning. So I would not recommend anybody to build an own LLM in a way but really to customize using LLM infrastructure which is like open source and commodity and then adapt it with your individual data. And there I think that's the real USP then that you have. And the outcome is that you have this combination. You have like you saved some time due to this like uh, uh what you do there and then really what is I think for the German Mittelstein coming back to one of the key advantages, this customer relationship. You have more time really to listen to your customer, to invest in this relationship, to deepen the relationship and the trust and there you have the help of tech enabled, whatever what you can do there. But it's really like yeah you have some more time to really think about or even think about the future. Think about new features. Um, in combination listen to your customer and then let's like adapt it and do some product and services iteration. So this human in the loop and then feeding your tech systems clear.

Speaker A: Um, so coming to, to to a point to the middle stand understands this how um could they really start? I mean they are now maybe approached by tons of consultants companies. Hey you need to do training for your employees. You need to work on predictive maintenance. You need to automate your processes. But now you need also to think about your, your business model. So there's a lot of people coming to the uh, CEO of a company and but what would be your advice how to start, uh, I mean with all the noise, how to see a clearly a way for pay, a way forward for your company.

Speaker B: I think very pragmatically but then also iterative and gradually. So um, identify one or two use cases for example and park the others and maybe take a use case where you, where the risk is not that high that it will fail because you have to convince also employees for example. So don't go always the hard way first also to learn, right? And then identify a use case is the first step. Um, then I think like less invalidating a uh, pilot, um, in some sub segments or something and then really think about the product, uh, and the rollout and then just then apply for a bigger budget. So really also start small with testing with a little budget. Um, and then really think about, yeah, ah, like rolling out. And as I mentioned later really first thing is the pilot, the testing, the validation and then think about like standardization, then the next like integration it and then also thinking about like the usage. Then it comes to the transformation. But of course um, the transformation then is the key for, for the success. And then the second thing which is where they start or what they should consider is like how's the government, who is taking decisions? Um, the next thing is how secure is our data. So where do we host the data? Where do we want this and how is this protected? And then the next thing, and we already talked about it like employee education because it's just that good how people use and apply it in their daily do work.

Speaker A: For me the question is still there. Um, with the, with AI we see a huge speed um, of how fast things are changing at the moment. So every week new news popping up. There's a new agent system here, a new capability there. So there's a massive change. So the speed is really increasing exponentially right now. And so we also know our German Mittelstand and they love to plan, they love to go through and make a strategy and make a plan and execute on the plan. Is there still room for a plan or uh, should it be? Uh yeah. So yes, we have now a five year strategy and now we can execute on the five year strategy. Is this still the time for that? Is it valid to have this long term plan or is it kind of outdated and maybe a problem in this AI era? Um, we are entering right now. What do you think?

Speaker B: So I don't know if the plan is the real answer. I think what. But what is really important and what I I realized that some companies lost it. It's like the North Star. So where do I want to navigate my company? What is my like a customer relationship? What is my real usp? And then I think you have to think about the way how to achieve this. Um but I think what is really important you have this long term vision. Especially in a family owned business. You have this long term vision and then you have to think about how you get there and what we do. Um, for example within Bridgemaker but also um. I advise a lot of companies that like um to have a like AI task uh force. So like really a few people who are really experts and who are like observing all this trending what you can exply and they also give them a budget to talk to experts for example and to be in this like not reinventing everything by themselves but talk to expert. Inviting experts to that you can always redefine the way or if there is a better systems. But as. As you said right. Sometimes one week it's this basis data uh lake and then uh the recommendation of using just a rack using existing database, not training them. It's also like yeah it's so many um variables uh which are changing there. And this is really like yeah I would recommend have like some people who are really trained in that and who are also like observing all these changes in the market.

Speaker A: Yeah, I also say yes uh, you need to have those kind of speedboats in the company exploring and testing out things. Um, allowing you to make failures um and to yeah maybe just to test things and learn from that and then you can transfer the learning to the other people in the company. So for me right now also the topic of venture building so kind of exploring becomes even even more important because for a long term vision you need to have a clear understanding what is my. My business model in the future and then I can transform into that.

Speaker B: Um.

Speaker A: So um. We also need to bit of close our podcast today. It's a fantastic conversation. Um and we always close with asking our guests about advisors. So what would be maybe a 1, 2, 3 advices you would like to give to the leaders listening to this podcast to the leaders of the German Mittelstand how to approach this new AI era?

Speaker B: Uh, I think the first is and that's what we mentioned at the beginning really think about automation is a necessity. It's a duty. So you have to do that. And then the beauty lies in beyond core. Right? So really think about how can I use AI to build better products and better services in an even faster, in an even faster way. That's, that's the first advice. And maybe then Innovation Horizons or other frameworks can help with that. But really think about like, what is beyond core. What is like in the long term differentiates me from all my competitors, for example. The second is, um, what I mentioned with experts, like people from the outside invite people who can really give you some new impulses where you can learn, where you can train, even like maybe in your advisory boards in your family, uh, businesses like lay some or add some, which are responsible for AI or innovation that you have like it in a very regular, uh, way. Third, keep an eye of your employees. Ah, what I meant with education and really applying, uh, AI tools, uh, in their daily day, because otherwise the real transformation won't, uh, be successful when it comes to AI transformation. And, uh, I think the fourth is, um, which is like my, my passion since I founded a bridge maker and still running BridgeMaker as CEO. It's like, really, how much time? Always ask yourself how much time are you thinking about the future? How much time are you investing in strategizing your future? And with AI now you have the opportunity that you get some extra time due to automation. You have some time, which is new. So really think about what and where do you invest this newfound time to make your company future? Proof.

Speaker A: Very, very good points resonates also very, very well with me. And I always also said to the leaders, um, they don't need to think too much about automation, but they can ask themselves a question like, what if you could have 100 or 1,000 more employees tomorrow for the price of one, what would you do? Um, so this, this kind of thinking of like leveraging AI, building something new because it's kind of cheap, um, but you can expand your capacity. But this is this kind of question nobody asked themselves so far. They always look at automation, but not really about what if I could do, if I could have more, more employees attend or whatever. And this opens the broader perspective.

Speaker B: And um, maybe, maybe this is very practical and I really like that. Maybe just have a look at your own calendar as a leader. So how much percent of my week? Uh, it's when do I think about the future? When do you think? Like, and AI, of course, applying AI technology is also part of that. Is maybe on Sunday evening or Monday morning really come to your calendar and how it looks like for the week and maybe then reshuffle a little bit Safety.

Speaker A: Very good points. Um thank you very much Henrike for joining. Was a fantastic conversation and very good insights.

Speaker B: Um um thank you so much for inviting.

Speaker A: Sure. And I just wanted to say uh. It doesn't feel like we are competitors. Feels like more we are partners and helping companies um becoming more um successful in the future. So thank you very much for joining.

Speaker B: I think that's what I maybe one last thing because it's really I absolutely we have to think about more about collaboration than competition in every company because um then we will have a change as Europe for example or German Mittelstone company really to think about more in ecosystems and collaboration. That's also your topic so you can't emphasize this enough really work together. Um yeah. And trust in this collaborations.

Speaker A: Perfect. And thank you for this. I keep it like this. Thank you very much for joining.

Speaker B: Thank you. Bye.