EP257: Building in Europe: Capital, Governance, and the Cost of Moving Fast
In this episode of The European Perspective, Jens Heitland is joined by Stephen Brooks and Annette van Berge Henegouwen to discuss the SAP and Prior Labs deal, European AI funding momentum, the contrasting governance paths of Volkswagen and Continental, and the hidden cost of unmeasured AI agent spending in large organizations.
Youtube Podcasts
What Europe's Latest Deals Reveal About Capital, Governance, and Speed
Working across large organizations for close to thirty years, I have learned to pay attention to the moments when two situations look similar on the surface but diverge underneath. This episode of The European Perspective, our third, gave us several of those moments. Stewart was away this time, so Stephen Brooks and Annette van Berge Henegouwen carried the conversation with me, and their perspectives shaped where the discussion went.
We opened with SAP's investment in Prior Labs, a UK-based tabular model company that recently attracted well over a billion in funding. On the surface, this looks like a European win. Underneath it, Stephen raised the more uncomfortable question. Microsoft is the one funding Mistral's infrastructure buildout, which means an American hyperscaler is paying for European capability while embedding it into its own cloud. The capital may be flowing into Europe, but the structural benefit is not automatically staying here.
That tension carried into the funding stories that followed. Humanoid, a two-year-old UK robotics company, closed a valuation north of two and a half billion. Cusp AI and Neko Health added hundreds of millions more between them. Annette pointed out that most of this capital is still coming from American hyperscalers and labs, which raises the same question from a different angle. Europe is producing valuable companies, but ownership of the growth that follows is not guaranteed.
Where the conversation moved into something more encouraging was entrepreneurial education. I have spent part of this week with the European Innovation Academy, watching three hundred students build startups from scratch over three weeks. Annette made the point that most education systems reward compliance rather than curiosity, and that the skills entrepreneurship actually requires- pitching, convincing, taking risks, are rarely taught directly. If European ecosystems want more founders who can compete globally without leaving the continent, the education system needs to catch up to what the market already rewards.
The clearest contrast of the episode came from two boards facing structurally similar pressure. Volkswagen, anchored deeply in German industrial and social structures, is trying to bring its entire organization along through a long co-creation process. Continental took a different route, selling ContiTech to a private equity firm and returning to its core business as a pure-play tire maker. Stephen made an important distinction here. Continental sold to clear its balance sheet, an option Volkswagen does not have. Annette framed it as a question of focus, knowing where you intend to play and refusing to be distracted from it. Neither approach is inherently correct, but they show governance responding to the same kind of pressure in opposite directions, and the outcomes over the next few years will say a great deal about which kind of pressure each company was actually under.
Banking added a similar layer. Berlin's agreement allowing UniCredit to take a stake in Commerzbank came with conditions that preserve the existing structure almost entirely, protect SME lending, a Frankfurt headquarters, and an independent listing. On the same day, the EU announced a banking reform blueprint aimed at reducing capital requirements to get lending flowing more easily across the continent. Annette summarized it well. Integration in theory, ring-fencing in practice. It is a pattern that shows up whenever European institutions try to move in coordinated directions while national interests continue to move independently.
We closed on something I have direct experience with. Several large corporates spent heavily on AI agents over the past year without measuring what the tokens were actually producing. I know this because I made the same mistake early on, burning three hundred euros in a single day with no outcome to show for it. What some organizations are doing now is shutting the agent programs down entirely rather than fixing the measurement problem underneath them. Stephen called it correctly: they didn't test before they invested. The lesson is not to abandon the technology. It is to measure what each agent actually costs against what it returns, something I now do deliberately, and something that turned my own token spend from hundreds of euros a day into a couple of euros for the same output.
None of these stories are unrelated. Capital flowing into Europe without staying in Europe. Governance responding to the same pressure in different directions. Institutions integrating in theory while protecting their own structure in practice. Organizations spending on technology before understanding what it costs to run. The throughline across all of it is the same question we keep returning to in this series: who actually benefits when something changes, and who is simply managing the appearance of change.
Human Innovation with Jens Heitland is produced by Heitland Media Group.*
Guest Links
Annette Van Berge Henegouwen: https://www.annettevanbergehenegouwen.com/
Stephen Brooks: https://www.linkedin.com/in/stephenabrooks1/
Website:https://www.jensheitland.com/
Business: https://www.heitlandmediagroup.com/
LinkedIn:https://www.linkedin.com/in/jensheitland/
Facebook:https://www.facebook.com/JensHeitlandofficial/
Instagram:https://www.instagram.com/jensheitland/
TikTok:https://www.tiktok.com/@jensheitland
X/Twitter:https://twitter.com/jensheitland
Newsletter:https://www.jensheitland.com/newsletter
YouTube:https://www.youtube.com/@jens-heitland
Subscribe and listen to The Jens Heitland Show Podcast here:
YT:https://www.youtube.com/channel/UCjuSGi1feauCNSER3IKuGWg
Web:https://www.jensheitland.com/podcasthome
Apple:https://podcasts.apple.com/us/podcast/the-jens-heitland-show-human-innovation/id1545043872?uo=4
Spotify:https://open.spotify.com/show/7H0GWMGVALyXnnmstYA1NL
Transcript
[00:00:00] Jens Heitland: Welcome to The European Perspective, the podcast that is talking about businesses and business in Europe. This time without Stuart, but with Stephen and as well Anette. Welcome back, everyone. Good to see all of you
[00:00:15] Stephen Brooks: Be back
[00:00:16] Annette van Berge Henegouwen: Wonderful to be
[00:00:17] Jens Heitland: After vacation, both have been on the water and enjoying the sun, and I was working and doing podcasts.
[00:00:28] Jens Heitland: Now let's get into it. So technology is number one, and there has been a lot of things happening over the last week and weeks in technology and specifically in AI.
[00:00:44] Jens Heitland: Let's go into the first topic. SAP paid over one billion for Prior Labs, which is a tabular foundation model layer.
[00:00:55] Jens Heitland: Maybe for everyone that has never heard about a tabular model, it is Excel instead of text inside of an AI model. So a large language model is learning from language and a tabular model learns from relationships in-inside of a table. That's maybe the simplest way to explain it. So from Steven, Anette, who wants to go first?
[00:01:25] Jens Heitland: What do you think about that deal?
[00:01:27] Stephen Brooks: I'm happy to dive in.
[00:01:30] Stephen Brooks: no pun intended, having been on the water for a holiday. I think it's easy to get swept away with the headlines of billions, in my opinion. But if we look at the underlying plumbing, SAB bought Prior Labs, but who's paying for Mistral's massive one gigabot computer target? It's Microsoft, an American hyperscaler paying a European lab to power its own cloud. that's quite amazing when you think about it. So while we're all compounding capability in region and the structural question remains, are we building independent European giants or are we just cultivating high-end local assets for foreign infrastructure to exploit?
[00:02:19] Stephen Brooks: What, what do you think, Aneta?
[00:02:22] Annette van Berge Henegouwen: Oh, that's an interesting question, Stephen. what I've noticed in the past week is that, beside the SAP deal, a lot of the investments are being done by the American hyperscalers and the AI labs. And it did make me wonder about the build we are doing in Europe to become more independent
[00:02:46] Stephen Brooks: Yeah
[00:02:46] Jens Heitland: what is interesting, double-clicking on that. if we take Paralabs, I think they're UK-based, so let's consider that Europe, Stephen. I think there are billions going into the company, so that's a good sign.
[00:03:09] Jens Heitland: If we look at Microsoft and them investing and working with Mistral, it's just a smart move. On one side, they're making themselves independent from their own investments, OpenAI where they still need to pay for the token.
[00:03:26] Jens Heitland: On the other side, they embed a local model into their infrastructure, which means they can run it for free in the future, I'm pretty sure. then the third level of that is regulation coming up now. So the AI regulation for Europe is going to be massive, changing a lot of things. I was even looking into, into that, what it means for our company, and I think we need to even change stuff for, for us.
[00:03:54] Jens Heitland: But even more important, GDPR. you have a local model, if you have it on servers in, in Europe, all of that will be easier to manage for them. So I think it's just smart investment and, and definitely helpful for Europe because we have a global company like Microsoft investing into a European company and, working with them on the commitment, that means all European organization will benefit from the intelligence and the growth of Mistral, which I think is, is a good, good sign.
[00:04:26] Jens Heitland: Other thoughts around it?
[00:04:27] Stephen Brooks: Yeah, listen, I, I, I think the default answer a- as to which AI infrastructure do we buy just changed this week anyway. And it, and it moved from a theoretical concept to an active procurement decision. So as you say, with Microsoft coming into play, I think people are gonna be looking at this a lot, lot more.
[00:04:46] Stephen Brooks: What do you think, Aneta?
[00:04:49] Annette van Berge Henegouwen: Very fair point, but also the on-prem, the developments of the QME3. there's so much going on, it's hard to keep up
[00:04:59] Jens Heitland: Yeah, it's huge. I wish I would have a huge server here in my in my office to test all of that stuff. I'm still dependent on tokens and small models.
[00:05:08] Jens Heitland: Let's get to the next topic. inside of technology, Humanite got a huge investment as well, and they're-- I think they're valued at US dollars of 1.3 billion, which is interesting if you look at a two-year-old company that is building robots, and they're UK-based, again, Steven,
[00:05:29] Stephen Brooks: Yeah
[00:05:29] Jens Heitland: your island
[00:05:32] Stephen Brooks: The mothership
[00:05:35] Jens Heitland: Now, what is interesting about that for me is it's again a signal for the European market that if you build something that generates global demand, you have possibilities. And you don't need to be in Silicon Valley, you don't need to be in the US, and I think that's something we need to push more in Europe so that European startups have the chance.
[00:05:59] Jens Heitland: And they got so much funding as well as part of it. for everyone digging into that, look up humanoid from the UK. It's super interesting, and I think it's a good signal founders and people building startups in the next coming years in Europe
[00:06:14] Stephen Brooks: Absolutely. Just two years old. based in Cambridge, 2.6 billion valuation. that's in two years. That's a hell of a valuation, and has attracted people like Bezos and Nvidia,
[00:06:30] Stephen Brooks: in its funding, which rubber stamps what they're doing and the way in which they're moving forward.
[00:06:39] Stephen Brooks: but I think it's interesting that the venture pipeline has matured for mid-cap boards, and the old rule that you have to flee to the US for a listing or a mega round, I honestly think is officially dead now
[00:06:54] Jens Heitland: Yeah, I agree. But I think it's still, I mentioned before we started recording, I'm supporting the European Innovation Academy this week and next week, where young students from all around the world are going through an accelerator building and testing themselves in building a startup.
[00:07:17] Jens Heitland: I think these are things that every school kid should be going through. How do you build a startup? How do you think entrepreneurial? And with that as well, how do you take risk in different ways? And if you ingrain that in the educational system so that the kids of the future all have that, I think one part we will build better companies.
[00:07:40] Jens Heitland: The other part is the whole ecosystem will grow up with that. And more and more people will look into how can I generate a company that creates value for other people solving problems. I think more companies in Europe that are established here and bringing revenue into the, economy and jobs with that will help all of us.
[00:08:04] Jens Heitland: So I think it's, it's, it's quite interesting to, to see all of these activities. We should just do more of it.
[00:08:10] Annette van Berge Henegouwen: I think the points you're making there, Jens, about is a very important one 'cause if we look in the UK, I think it was Google's DeepMind was there. There's this place in London, and Steven, you might be able to elaborate, where been a lot of activity, early activity, uh, in the tech and, and now it's a blossoming ecosystem.
[00:08:32] Annette van Berge Henegouwen: So building ecosystems is actually a very important, uh, aspect of being successful
[00:08:38] Stephen Brooks: absolutely right. Jens, I'm interested, and Anette, a question for you both on this. Do you think you can teach entrepreneurialship? do you think it comes from within to start with, and then you can educate and share the right way to go about it? All entrepreneurs have to fail, and I wonder if education meets that gap between education and business, and then education and entrepreneurialship, and then entrepreneurialship and growth
[00:09:15] Jens Heitland: Yeah, I think we need to reinvent the education in general. That's a longer topic. But when it comes to entrepreneurship, it's not-- I think it's not about education in the old school way. It's learning by doing and then having experienced mentors and people that have failed a lot, helping them to not make all the stupid mistakes and learning from the mistakes, but also helping them, and that was my role this week, helping them to ask the right question at the right time.
[00:09:47] Jens Heitland: Because as a mentor that has ... I mean, I'm double the age than the students right now. I have quite a lot of experience, and I have failed a lot in my, in, in, in the history. Just helping them on certain things. Like, "Hey, I love your idea, but what problem are you solving?" If we take that example. So just getting them, hey, nail the problem first before you build the solution, so that you have the person that has the problem, the company that has the problem, and then from that you build the solution and you fine-tune the solution.
[00:10:20] Jens Heitland: But you do that iteratively. It's not you, you, you assume and then you build and then you sell. It's like it's very, very fast-paced, especially with AI and digital development these days. But I think to your point is you need to do that in a way that they experience it. So it's, it's an experience, and that's what, what, what I love about the, in, uh, European Innovation Academy.
[00:10:44] Jens Heitland: They do this as an experience with next week it's 300 students in, in one experience, and they all do different startups, and they all help each other. Uh, and they do that for three weeks, which is an amazing experience for them.
[00:10:59] Annette van Berge Henegouwen: and to that point, it's the Innovation Academy, so it's learning things. It's learning skills. If we start earlier, and some schools do that, some universities do that to stimulate curiosity instead of stimulate to be compliant and sit still and do as you are told. no entrepreneur got where they were by being complacent and sitting still and doing what they were told. So I think the education system has a large role in that. But it's also hardcore skill building. How do you do a good pitch? How are you convincing in delivering a message? You can be a great entrepreneur, but if you are incapable to convince any investor of the value of your idea, you're not gonna get there
[00:11:53] Jens Heitland: Yeah, agree. I think it's unpacking all of the skills and all of the things that are baseline. we have mathematics and languages. I think we need to break down business and entrepreneurship into similar things and then understand what are the foundational principles, and then from that on, how can you, through experiential learning, build an educational system that allows kids from a young age to get into.
[00:12:23] Jens Heitland: Maybe that's something we should build a business around, but we can do that after the recording. Let's get into success stories. So we already talked about humanoid. Let's, let's look at Cusp AI. Again, two point six closed in four hun- no, they closed four hundred and fifty million. All again, uh, the UK. then Neko Health closed US dollars seven hundred million on the 15th of July.
[00:12:52] Jens Heitland: Shows there's so many things happening in different industries. So we had robots, we have had technology again, and then we have healthcare. So I think there's a lot going on when it comes to funding, a lot going on when it comes to investments, especially in the first quarter this year or the first half of the year.
[00:13:16] Stephen Brooks: Absolutely. Europe is proving it can catch world-class assets right now. I know we're still outsourcing the heaviest of growth to Silicon Valley, but for those entrepreneurs we were just talking about for the future, the funding is there. People are looking at Europe with anticipation, and some of these success stories I think bear that out
[00:13:44] Jens Heitland: Yeah, I agree. The interesting part of that is it's not just the international organization investing into European startups, it's companies like Schaeffler and Bosch who are investing and doing deals with European startups or grown-ups, scale-ups, which I think is also a signal for other European companies to invest more into European startup ecosystem.
[00:14:11] Jens Heitland: And maybe instead of going to one of the obvious when you do a procurement process where you say, "Ah, we always work with the big five companies in this sector," why not looking into is there a European player that could deliver the same, but they don't have the big brand name like Microsoft or whatever it is?
[00:14:31] Jens Heitland: I think it's worth considering in big organizations now as well, because with that, you're strengthening Europe and you're strengthening the ecosystem again, like we said before.
[00:14:42] Stephen Brooks: Thijmen
[00:14:43] Annette van Berge Henegouwen: Hmm. That's an interesting point you're triggering there, Jens, 'cause before we talked about having excess capacity to be able to withstand geopolitically induced difficulties that one might have. Steven, as an entrepreneur, how would you be looking at creating, buying, procuring additional capacity, but not just from a big reliable source, but also from a smaller source so you have more capacity but are sourcing local,
[00:15:25] Stephen Brooks: I, I
[00:15:25] Annette van Berge Henegouwen: even from the UK?
[00:15:26] Stephen Brooks: I think sourcing locally doesn't have to be local. I think you've got to look for people that are in the niche that you need expertise or product or services from, and then that resource from them. Listen, I work with people outside of the UK all the time, across Europe and the Middle East and India and America. But that, you're finding people that are just an expert in their field or they produce that one widget that you need that makes your machine work. And I think that's where this local buying comes into force. If we think of it's counterintuitive, but if you think of local being global, finding the niche of the person that you need, I think that's what we should be doing.
[00:16:25] Stephen Brooks: I'm all in favor of let's find it in Europe first, and if we can't find it, then we have to go further afield. But as we've seen, listen, all this funding isn't just a lucky one-off round of funding at all. It's a deep tech stack funded by a strategic industrial capital that we don't need to go to the US for anymore.
[00:16:46] Stephen Brooks: So let's find those local businesses. Let's find those local entrepreneurs, Jens, that you're encouraging and let's support them and grow all of our businesses in that way
[00:17:00] Jens Heitland: Let's go to the next topic, transformation and change. Two boards, same continent, opposite governance outcomes. have Volkswagen. Let's start with Volkswagen. Yeah, I think in the end, and it's always easy to say from the outside that's why we do it. if we take Volkswagen, they're anchored in German society, and I think they just need to into how do they co-create, and that's an easier word said than done. But it's looking into how do you bring all employees as part of this, external parties are in the board into it and develop it together.
[00:17:50] Jens Heitland: As difficult as it is, I think it's all about how do you as a global but German organization it to a level that you have conversations and most probably that's a long-term process that needs to happen. But everyone understanding the impact, everyone understanding the problem, everyone understanding in depth what need to be changed, and then a momentum on that co-creation to then build the future together.
[00:18:22] Jens Heitland: That's my two cents on that. I think it's super difficult, especially, I'm German, I know how difficult it is in Germany. and they're still very old school, I believe from an organizational context and leadership structures and everything. So it's not going to be easy, but for me it's that need to happen somehow.
[00:18:48] Jens Heitland: Steve, what, what are your thoughts on Volkswagen?
[00:18:51] Stephen Brooks: I think the key here might be communication. the vision I don't think maybe ... And we don't know, right? Let's all be clear on this. We don't know the politics of the VW board or its advisory board. I mean, nobody wants to see jobs at all. That's a given. But I think, if you're creating a North Star and you're creating a vision, you've got to be able to express and communicate that vision and take everybody along with you on the journey.
[00:19:25] Stephen Brooks: And it's just we were saying about the people in the entrepreneurial innovation. don't have that skill or it's not a skill that is polished as the leader, you're gonna struggle, right?
[00:19:42] Stephen Brooks: in times of crisis, whether it's a government, a company, one looks to its leaders for guidance, leadership, and strength. And if you can't communicate that, that's when it starts to fall apart. and I think if the circumstances were different, maybe everybody should have got together right at the very start to say, "What does this plan look like? Where do we take that business?"
[00:20:14] Stephen Brooks: Yes, they're struggling in China, but look what's happening in the Chinese motor industry with BYD and other businesses like that. they're expanding across the world and even taking market share away from Tesla and other brands. So VW, I think, might just have to stop, take a breath, and go, "Where else can we innovate?
[00:20:38] Stephen Brooks: What are we good at that we can compete with in the market?" And that, might be you know, selling parts to other businesses. It might be sharing design. It could ... I mean, there's all sorts of things in the motor industry, just as there is in any other industry, isn't there? But, I think the heart of this problem comes down to communication, and, not communicating that vision effectively enough that you bring people on the journey with you
[00:21:08] Jens Heitland: let's go to the parallel example. Continental sold ContiTech to Lone Star for four billion euro. And what is interesting is they go back to be a pure play tire maker from September on, it's all about how do you-- almost going back to the roots, I guess, for them. it also shows they have created massive value with something that they did, which this with the ContiTech, then selling that so that you can concentrate on the let's say maybe on the basics again, and something that you have done over a lot of years and then go back to the roots.
[00:21:57] Jens Heitland: I think it is an interesting play, and it looks like it works for them, if we compare that to Volkswagen at least, and they're still in the car manufacturing or let's say wider automotive ecosystem as well. Thoughts around that?
[00:22:14] Annette van Berge Henegouwen (2): What comes to my mind is the importance of focus and choosing where you are going to play and not be distracted from that. And it could be a vertical integration, it could be something horizontal, or it could just be a pure tire play. but the importance of focus and knowing who you are and where you wanna go, that's what comes to my mind.
[00:22:43] Stephen Brooks: I'm reminded of a phrase of a guy I sat to on a plane 20-odd years ago, and it was, "Know what you know, what you don't know, and be known for what you know." And Continental is known for tires. and they've gone back to the core of their business. Let's not get confused, though, these are two structurally identical problems, situations. Continental did sell out to a private equity firm, Lodestar, to clear its balance sheet. Volkswagen don't have that opportunity. So I think it's... Listen, it's a great repositioning for them. and given that everybody needs tires on a car, I think their business is a lot safer than perhaps Volkswagen is, but my hat goes off to them. Great work
[00:23:46] Jens Heitland: Yeah, I agree.
[00:23:46] Jens Heitland: Let's get to the next topic. Industry deep dive. Banking sounds financial, Anette. So you have been--
[00:23:58] Annette van Berge Henegouwen (2): doesn't it?
[00:23:59] Jens Heitland: Yeah, you have been diving into that already. Share. what, w- what is, what is the topic in banking this week?
[00:24:07] Annette van Berge Henegouwen (2): It's interesting. something shifted in Berlin, after quite some difficulties, Merz agreed to UniCredit taking over Commerzbank, but not just allowing the takeover, but at a price, and it's keep lending to the Mittelstands SME book, the headquarters in Frankfurt and keep Commerzbank independently listed. So it's you can invest, but you cannot touch any of the existing structures. And in regulated industry in Europe, if you don't touch the structures, then capital demands are quite high. At the same time, the EU banking reform, a blueprint was announced where in line with von der Leyen's line, getting capital flowing is how we get Europe growing. They want to reduce the capital demands and look at it from a headquarters perspective and not from a subsidiary perspective for the amount of capital that needs to be obtained, and therefore increasing the lending possibilities. But this was happening on the same day. So Berlin creating conditions, ring-fencing in practice, whilst is saying, "Let's integrate further," but integration in theory and ring-fencing in practice
[00:26:02] Jens Heitland: Yeah
[00:26:02] Annette van Berge Henegouwen (2): So not, not, not a, not a convincing situation
[00:26:06] Jens Heitland: I agree. I think that shows again that we're not all playing together if we take Europe, and I think that's specifically in old industries even more so, but also politically influenced obviously. I think that's something we need to get our things together in Europe where we say, "Let's work together on things from the get-go and align for going to happen and so that these kind of things are not working against each other."
[00:26:42] Jens Heitland: Um, Steven, thoughts on this?
[00:26:45] Stephen Brooks: a love-hate relationship with banks. Like most businesses that they're there when you don't need them, and when you do need them, they're absent. but I looked at this with some interest and looked at the fine print on these Berlin negotiations, and they haven't wavered the UniCredit, but they have laid down three heavy defensive conditions. Preserve SME lending, keep the HQ in Frankfurt, and maintain an independent Commerzbank listing. Plus the ECB approval clock is a brutal two-year process, right? So this isn't a free market green light. It's a political compromise forced by the fact that Andrea Orcel maneuvered them into an equity corner.
[00:27:37] Stephen Brooks: Capital might flow easier under the, the commission rules, but national politics will still fight cross-border banking tooth and nail
[00:27:45] Jens Heitland: Let's get to the next topic, entrepreneurial thinking, to get us out of the banking and politics. Step. It looks like Ben Francis is buying back his percentage in Gymshark stakes. So he was the founder, which I think is interesting and I see this more happening that founders go back to take the business that they already left and lifting it up again.
[00:28:20] Jens Heitland: Like almost Steve Jobs did with Apple the, in, in the old days. We have seen it as well in, in the US where I think it was earlier this year or, or end of last year, where the Google founders went back into operation and said, "Let's, let's, let's tackle, tackle certain things." I think it's interesting that, people like them doing that because they definitely don't do it for the money, I'm pretty sure.
[00:28:50] Jens Heitland: Stephen?
[00:28:51] Stephen Brooks: Listen, I've got personal experience of this. I exited a business a number of years ago, and then went back five or six years later because it was broken and needed fixing. And I'm not saying this is the case with Gymshark, but it is classic founder entrepreneur mentality that you look at the business you created and you see with hindsight after you exit, "Hang on a sec, there's a better way that we can be doing this.
[00:29:23] Stephen Brooks: This is, this is what the n- new innovation in the business should be. This is what we should be trying next." they don't see the business doing it, so, you know, just like jobs, just like myself, y- you wanna get back in there and, and shake it up and then, and then guide the business itself. I take, take my hat off to Ben.
[00:29:41] Stephen Brooks: I, I think it's a, it's a great move. a- a- and we'll, we'll look with interest to see, to see where this business goes under his, under his guidance again.
[00:29:51] Jens Heitland: Then o- other success stories from a, from an entrepreneurial perspective, we talked already about a couple of them. A, a lot of funding goes in, in, into, uh, startups. Munich Space startup as one of the, examples with, with 10.2 millions. But if we talk about entrepreneurial thinking, one of the things that I've experienced over the last couple of weeks is that a lot of large global organizations that are headquartered in Europe are rethinking how they're organizing themselves because what they find out, is they don't really innovate anymore.
[00:30:37] Jens Heitland: what they're looking-- and it's always big transformation that are happening, and I hear it from the inside of that organization, so it's not public data. But all of them struggle. How do you get an organization that is so big to innovate from within and getting this entrepreneurial thinking?
[00:30:57] Jens Heitland: everyone talks about it, but a lot of s- companies struggle with that. Stephen, you have been working with big and small companies plus owning. What do you think large organization need to get, get into over the next, let's say, year to figure out their innovation pipeline and, like, getting entrepreneurial thinking back?
[00:31:20] Stephen Brooks: Yeah, and taking away the hub of the business and moving it somewhere else or disseminating the headquarters of a building isn't such a bad idea. In fact, the new prime minister here in England, Andy Burnham, is gonna have a number 10 in the north, right?
[00:31:42] Stephen Brooks: So he's gonna be moving some of the power to the north of the country. I think there's a real issue between people understanding transformation and innovation. And people have got to get uncomfortable with the fact that innovation involves risk.
[00:32:06] Stephen Brooks: Innovation involves talking to people that perhaps you wouldn't even include in a discussion around what do we do with the business? How can we make it better? What can we be doing, right? And it's only by having those conversations and moving that forward with people that aren't generally in the decision-making process, that you come up with great ideas and new ways of looking at things or new directions that you can take the business. Transformation is very different. Transformation is about how do we take what we're doing into something a bit more manageable? maybe it's smaller, maybe it's bigger. that's not innovation in my opinion. So I think, devolving that out to everybody in the business, just Volkswagen perhaps should have done, everybody's got a voice, right? and then filtering that down into some kind of innovation process, innovation hub within the business that can sift through those ideas and work on the ones that are gonna change the business, and have somebody that's looking at transformation of the business in a total separate silo, as far as what's going on
[00:33:32] Annette van Berge Henegouwen (2): Yeah, super agree with you there, Stephen. Transformation, anything that is within an existing business, within the existing organization is about optimizing. You can make it better, but you cannot fundamentally change it. Any system exists on its own because it's working to some extent for the stakeholders in that system. Even if it's not working, it's working on some layer it's working. if you look at innovation, what you need is something separate, an ambidextrous organization, something outside of the overall system where there are not the powers are, so to say. To be able to do things differently, to look at reality differently, you need to be in a different reality, and that's not as a part of the primary organization
[00:34:27] Stephen Brooks: Absolutely agree
[00:34:28] Jens Heitland: Let's get us to the last topic, F-up of the week. One of the big... And you
[00:34:36] Annette van Berge Henegouwen (2): Jens.
[00:34:37] Jens Heitland: go first. You go first, Aneta. Now you-
[00:34:39] Annette van Berge Henegouwen (2): I go first. with my frozen image. it's clearly my setup that's not working, so yeah. I'm calling dibs on the fuck up.
[00:34:55] Jens Heitland: Yeah. So for those that haven't realized that are listening to the audio only, Annette's image is frozen most of the time. And she's at home, so she's not at the boat like we planned in the last weeks.
[00:35:12] Annette van Berge Henegouwen (2): Yeah. But back to you, Jens, please
[00:35:14] Jens Heitland: Okay, let's talk about F-Ops. again, big corporates in Europe, and I cannot name them. I should not name them because I have heard about it behind the scenes, but I have heard it from three already. What is happening, everyone was racing in token maxing, however you want to call it, over the last half year.
[00:35:38] Jens Heitland: Let's use AI out how it works, which I think everyone should do still going forward. what nobody looked at how much tokens are you burning, and with that, if you use the external models, they're quite expensive. And I can just give my example. I have built an agent system early, I think it was March this year, and I was looking into utilizing AI as much as I can, building agents.
[00:36:09] Jens Heitland: And if agents are working twenty-four/seven, then you increase quite a lot of tokens, worst case, and if you use the most expensive model, then even worse th-than that. So I had, uh, on one day I had a three hundred euro bill for just AI tokens and zero outcome from that three hundred, and that's happening right now in corporate.
[00:36:30] Jens Heitland: They have spent some of the millions of euros in tokens and AI agents and so on, and now it goes backwards. So the financial team has analyzed that, and they understand that there's something wrong because they have the cost, and the cost were there and verified or approved in an internal context to be able to generate new revenue, but the revenue is not coming.
[00:36:58] Jens Heitland: The upside is not coming, but the millions are spent. So what they are now doing, at least in some organization, they are killing all the different agent programs and bringing them back to basics to say, "Okay, this is not how it's going to work and you're not playing with agents anymore."
[00:37:18] Jens Heitland: Which I think is counterproductive because they should work with agents. It's just you need to understand how to do it. Today I'm spending two euros my agents run twenty-four hours, so it is possible. If I can do it, big organization can do it as well. Thoughts from you guys?
[00:37:36] Annette van Berge Henegouwen (2): Well, if I may add to that, Jens, uh, the token maxing was actually one of the KPIs that has been used in
[00:37:44] Jens Heitland: Yeah
[00:37:45] Annette van Berge Henegouwen (2): And with KPIs, designing KPIs, you always have to be conscious of the downside. And of course, token maxing is a downside in itself. So, that proved to be not the right KPI. But then the question is: What is the right KPI? How do you stimulate innovation? How do you stimulate learning? how do you stimulate people getting out of their comfort zone and doing new things having enormous token bills? Steven, any thoughts?
[00:38:26] Stephen Brooks: i think this is classic didn't test before invested, right? And the fact that people are killing off agent programs all over the place is for that reason. They didn't... Everybody just ran towards the shiny light, I'm afraid, in my opinion. And as... my background, as everybody knows, is pop-up retailing, and one of the things that we teach people in that industry is that you test before you invest.
[00:38:56] Stephen Brooks: You make sure that the ROI is gonna be there. Might not be as big as you expect to start off with, but there's an ROI from it. Jens' classic example, right? It's 300 euros, nothing in the pot, For some businesses, half a million dollars, nothing in the pot. That's half a million off your bottom line, guys, straightaway. classic you didn't test before you invest. Learn the lesson, go back, rethink, and again, it comes back to that innovation. Let's have people around that have got better ideas about how to use this technology in a way that will give you a great return on investment
[00:39:38] Annette van Berge Henegouwen (2): I think there's also in distinguishing between learning and budgets for learning and budgets for experimenting versus investing. I want to have people learning, but if you start changing processes, et cetera, you wanna do it somewhere where there's the shortest learning loop, so preferably high volumes, shortest learning loop, and then focus on things that add value.
[00:40:08] Annette van Berge Henegouwen (2): So take away, bottlenecks, uh, decrease throughput times, things like that, where you can also actually measure value very quickly. But the main point for me is distinguish between learning budget and investments
[00:40:23] Stephen Brooks: Yeah. Good point, Annette
[00:40:25] Jens Heitland: Yeah, I, I
[00:40:26] Annette van Berge Henegouwen (2): pro-- and of course, proper financial control. Sorry, Jens, over
[00:40:29] Jens Heitland: Yeah, that's over and out. No, I, I agree. It's, it's key part, and that's, that's literally if we go back to my, my small s-scale example, that's what I did. I built a measuring algorithm that how much tokens we use for what, but also what's the outcome of each of that. like a return of invest calculation, that scientific, but it works.
[00:40:55] Jens Heitland: And what... And I think that's, that's, that's... I think it's common knowledge, but I think it's not yet. What, what you understand if you dig deeper and if you work with this AI agentic topics every day, then you understand you don't need to have the professor, like the most expensive reasoning model for every single job.
[00:41:17] Jens Heitland: You can use a very, very small, cheaper or even free model for certain things, and you-- it, it just does the job as good as the other. But we always think we need to... Because if you look at consumer AI and you use, let's say, ChatGPT or Claude or whatever, you're used to this chat window, and then you want to have...
[00:41:41] Jens Heitland: Like, you have a subscription which is 20 euros, and then you want to have the best experience that you can have, and then you always choose the highest possible. But if you look at a-agentic AI, it's really all about what is the ag-agent's job, what d- what is, what is required to be done, and do you need reasoning for that high reasoning, low reasoning, whatever it is, and then use the specific model for that use case, like we have discussed in, in one of the previous episodes.
[00:42:12] Jens Heitland: I think it's critical to understand that as an organization. Unfortunately, for people like me, you only learn it by doing, and that's maybe going back to an organizational setup as well. you want to train and get everyone into AI and understanding how to use it, you need to invest time into that and build it experiential, because it's not that difficult, and there are tons of companies that are happy to help organizations to do that.
[00:42:44] Jens Heitland: But I think it's a learning system that you need to build out how you can work with AI efficiently in an organization so that everyone can leverage it.
[00:42:54] Stephen Brooks: I, I think there's also a case, Jens, of lots of businesses continually playing catch up
[00:43:00] Jens Heitland: Yeah is moving at such a pace that you just find something that just about works for you, and then there's something, "Oh, hang on a sec. We need to try this because this looks better." And, and there is that, that danger.
[00:43:13] Stephen Brooks: And I think Annette is absolutely right about having this learning budget, this com siloed budget where we can just invest time and invest money in learning what is best and then innovating with that, rather than just like I said earlier, running towards the shiny light 'cause that's what everybody else is doing
[00:43:35] Jens Heitland: Yeah, 100%.
[00:43:36] Annette van Berge Henegouwen (2): Ja, das war's
[00:43:38] Jens Heitland: Anette, any, any thoughts on this further?
[00:43:42] Annette van Berge Henegouwen (2): No, in agreement with you both. So
[00:43:46] Jens Heitland: Good. W- w- with this we have covered all topics today. Any- anything we missed? Except Stuart, obviously.
[00:43:54] Annette van Berge Henegouwen (2): Stuart
[00:43:55] Stephen Brooks: we missed you.
[00:43:56] Jens Heitland: No, no good. W- with that, happy Friday evening. Thank you for recording today again. Was a pleasure as always