Transcription
Thank you. Good afternoon. Um [snorts] my name is uh as we just heard Bonibini. I'm from Hamburg and um I've been working on the intersection of digital innovation and marketing business all kind of stuff uh for like over 20 years now. And my imaginary job title is as I always like to say practical visionary. Yeah.
I'm trying to find new trends before they really appear in the mainstream. I try to understand them then do a lot like like keynotes and speeches um uh about them but I also help my clients actually implement them. Yeah. And I that's the practical part and that also helps me then again to more understand these these trends and um I don't do that much in e-commerce. I'm more in automotive of telecom but through all the stuff that's happening right now and more and more I also do stuff in in in retail. Uh and as we just heard I'm a co-founder at the XR lab uh at the University of Müster but that's a totally different topic.
Uh what we're talking about today is a course that I teach at UC Davis uh in in California. It's called rethinking digital and that's also a workshop series here in uh in Germany. And the basic idea is how can we rethink digital with concepts from digital China.
Um me myself I was always going to the US to find new trends to find new interesting things. Um but like 10 years ago I realized that there's not much coming out of the US lately. Yeah. AI changed this a little bit. But if you go to China you realize that there are much more interesting things happening there than in Silicon Valley these days. And why especially China? It's because that's the only place worldwide um that's not influenced by Silicon Valley. Yeah. Because of for political reasons they have this digital firewall. Um but basically that means that um it's an ecosystem that no really doesn't get much influence from Silicon Valley in the direct way like everybody else on the globe does. And if you're in a different ecosystem, if you have different experiences, that's the place where you actually find different ideas. And that's why I think China is so interesting. Usually we don't look to China and that's why we very often get surprised by things coming out of China. Yeah.
Here are a couple of surprises of the last couple of years. Deep Seek last year. Yeah. Electric cars, Teu. Um all these are um surprises that we saw. Here's a rather recent surprise. Who saw this last week? Yeah. Get off the >> It is not Kane West. [music] >> It's a nice music video um that actually was generated by a tool that calls uh Cance AI. Who who saw Cance AI last week? A few more. Yeah, that was the first surprise of 2026. Uh I think and it comes from Bite Dance. I think we all know what Bite Dance is. It's uh like we have Tik Tok here from Bite Dance, but it's a much bigger company. What we don't know probably is that Bite Dance is actually in terms of usage the biggest AI company worldwide. Yeah, 50 trillion tokens every day are used. You see here how much Google AI, Open AI, Anthropic is using. um it's smaller than bite dance and nobody knows this. Yeah. And the interesting thing is also Bance is also an e-commerce company and how much revenue might a social network like Mike Dance do? I think uh Amazon does like 830 or something billion worldwide GMV by dance does 670 billion quite a high number and again uh very often when I talk about this another surprise and the interesting thing about bidance is it's not a social network it's not an e-commerce company it's not a company that provides um by the way you get all the slides afterwards so you don't have to take pictures. Um, it's also a company that does AI technology or productivity suite. Yeah, it it's all of the above. And that's something that is looks pretty totally alien to us. And that's the reason why China is so interesting. Yeah. Because it's a kind of time machine where we actually can look into our own uh digital future.
And now we're come back to AI in the west. Yeah. If you want to visualize what's happening with AI in the west, um, I asked an AI and they came up with this. Yeah. Um, it's a data center with a lot of money burning in it. Well, that's probably a good picture uh to have in mind when we talk about AI. I think you all know the numbers. Um, it's economy level investments that they are taking hundreds of billions every year. They plan to do this for another five, six years. Unbelievable. Yeah. And um if you look at the economic side of it, it's actually much bigger than the.com bubble or the subprime bubble in 2008. And we all know what happened in 2008 when it burst. So it will be very interesting to see uh what hap what what will happen to the AI bubble. Uh and what do we do with all this money? What are all these tech companies are trying to achieve? AGI, yeah, artificial general intelligence, a kind of super intelligent that then rules the world and justifies all the money that we pour into this. That's the wild goose chase they are all on. What do we get until we're waiting for this AGI? Mostly bold ideas and cool demos. Yeah. And cool demos very often look like this. Yeah. And very often all the stuff that we are doing with AI also only gets stuck in the demo version uh uh phase like for example Clana who wanted to fire all their service people uh two years ago because AI can handle that. Yeah. 6 months later they hired them back because they realized uh the proof of concept work great. Yeah, the pilot work great but all the edge cases makes it extremely difficult to actually do this. Yeah, you actually need to have the the people back and um well that a great example I think for demo versus uh scaled and I think you're all familiar with these numbers and that's just one uh an old study from last summer but there are others from um PWC uh Boston Consulting Group had another one. It's always the same numbers. Yeah. It always says every company nowadays does something with AI, but very rarely we see really economic value coming out of this. Not on a small scale, but on a large scale. And that's a very common problem that we have there.
That doesn't mean that everybody doesn't like every speaker on the stage is talking about B AI first. Yeah. And every time I see this, I ask myself, okay, what does it mean? Yeah, very often I don't hear this. I don't hear any real concrete example for what does AI mean? And I have a good example that's from my speech last year. Shien, I think here everybody nowadays knows who Shien is. Shien is basically an AI first company. And the interesting thing here is they are not using AI as a tool, but they are using it as a platform. Yeah, we don't don't go into this detail. You can maybe still watch my keynote from last year. Uh we talk about Shien. Um that that's a very good example. The interesting thing is nobody's talking about Chien as an AI first company. But it's a good example for where the differences are between the West and China when it comes to AI. Yeah, we see AI as a tool that helps make existing processes more efficient. Yeah. And it's always about the bottom line. It's about savings. It's about firing a couple of people and saving the money. Um, in China, you use AI as a platform to actually rethink business models, come up with totally new business models like the one from Shien. It's always about that top line. It's about growth. It's about uh getting more revenue and and that's a very different approach, right? And in China, that also means that you have a different focus. you don't chase AGI, you the goal is AI diffusion, making sure that uh AI is scaled uh industrially. Yeah, really AI deployment everywhere and uh as I said, I'm more familiar with the automotive uh uh uh sector and I'm not sure you probably all heard about Xiaomi, the third biggest um uh smartphone brand. And the interesting thing is they announced that they want to go into electric cars and 1,123 days later they launched and actually start selling the SU7. Yeah. The the first electric car. It's an amazing car. Really interesting. And they are actually using AI to produce it. And that looks then like this. [music] And what what we are seeing here is what is called in China a dark factory. And it's so dark because there are no people in there. Yeah. There are stuff going in and there are cars coming out and in between there's a fully automated autonomous production uh of these cars. Yeah. And that's AI first. Yeah. It's not about the [ __ ] that we often talk about here. It's about being able to produce 40 cars per hour. autonomously. Yeah. And that's what you should start thinking about when you think about AI. Think bigger.
Um, but we are talking about e-commerce, right? So, do I have also interesting cases for uh AI? Sure. [snorts] Livestream is a very interesting thing that we have been talking about when we talk about China and e-commerce. And what you're seeing here is the founder of JD.com, yeah, the biggest retailer in China, uh, doing a live stream. The funny thing here is it's not him, it's a digital twin. Yeah, it's an AI agent selling stuff and uh that's already two years old. Yeah, that's how the whole thing started. Now JD has a service called Joy Streamereamer and they are actually virtual hosts for live streams and they are actually selling stuff. Now you see the number up there 2.3 billion R&B. It's like a quarter billion uh dollars in sales that these virtual hosts did uh last uh double1 season. Yeah. 11 of of November. Uh and the thing funny thing is 80 they they are more successful than 80% of their um uh human uh uh hosts. So pretty interesting thing. And that's really an application that generates real revenue. Yeah. It's not a test. It's something that is really happening.
A very another very interesting case that here nobody talks about that but I think we should talk about is AQ. Yeah, AQ solves a problem that uh here probably is not much really of a topic. If I have something to sell, yeah, there are lots of solutions to do this and they are actually doing lots of uh conferences and you have a booth out there for every service that helps me selling stuff. The interesting thing is where do I get stuff to sell and that's where ACU comes in. Yeah. AQ actually helps me is a AI chatbot that builds me supply chains. Yeah. Like for example that's a real life example. um somebody wanted to have sell Christmas slippers uh last Christmas season and uh that's the prompt that uh he put into AO [snorts] and then AO does uh market research helps you what could be a good design for Christmas slippers this season then it generates pictures yeah suggestions um based on those research what kind of uh slippers should you sell and here comes the interesting part and And it actually helps you to connect you to suppliers who can uh produce these designs that you have there. Pretty amazing. And the interesting thing is when you look to China, they are always trying to get the details right. Like for example, if you right, you see this little chat button here underneath the supplier. You can chat with the supplier. You chat with him in German or in English. He sees Mandarin, replies in Mandarin and you again see uh sees uh see see the answer in his language in your language and then you get an agreement afterwards. Yeah, that's comes from um the AI comes from Alibaba and where you see what you have agreed to and the interesting thing is these kind of AI translations is something that everybody here also does. Yeah. Google has translations forever. Depot has great translations, but the detail here is that Alibaba guarantees [snorts] that the translation actually works for you. Yeah. If there's a mistake in there, um, then it's Alibaba's fault. And that makes sure that you actually use this for this kind of business transaction. Yeah. The the task here that Alibaba sees is they have to scale the trust that we have in these systems. So they can actually be used for live business. Yeah. And then if you use this then you actually have a um a chatbot and AI that creates physical products instead of just nice pictures.
And I think you all heard about Alibaba, but you I'm not sure if you're all familiar with the ecosystem around Alibaba. Yeah, it's they have a lot of e-commerce companies. They also have a lot of uh cloud and AI operations, media and all these other things that you see here on this chart. And you can do on Alibaba Qen something that we probably have heard today also for a couple of times. You can ask your AI to do something like book me a cool hotel in Hoou and the train to get there and then buy me an umbrella or a hat depending on the weather. Yeah, that's something that we also talk about that AI AI agents at a certain point will be able to achieve. The funny thing is Qin can do this today. Yeah. because they have the ecosystem of all the different companies like for example Flegy a travel agent they have a map service that provides the weather information and they have lots of e-commerce that buys me the hat or the umbrella but they also have the technical infrastructure to do this. Yeah. And one of the things is they can they have to handle the um the uh uh payment and that comes us to the second part brings us to the second part of getting things right. Um, they have tokenized mandates. You can make sure that when your AI agent goes out and buys something for you, there are clear guard rails that make sure that it doesn't come back with a I don't know a car or something. Yeah, it comes back with the one the stuff that you actually want and you need this to have again trust into the system to actually use it and people are actually using this. Yeah, this Q1 agent can be used by more than 100 million people um today. Yeah, it's not a test. It's something that's really live and get actually used and they they call it a one-s sentence purchase. Yeah, though I had two sentences but uh you know what I mean. Yeah. And that's I think pretty interesting.
What can we learn from this? That we have to get or find a way to get from demos to dollars. Yeah. We are still very often uh stuck in this AI sandbox or proof of concept or pilot thing. Um, but to really unlock the value and the potential that's in there, we have to find ways to actually get there. And um that means for example, get the details right. Yeah. And here are four lessons that I think we can draw from this uh no [ __ ] approach. The first thing is aim for growth and revenue. Yeah. Cost savings are nice. do the cost savings perfectly fine but you don't grow by saving on costs. Yeah, that's pretty simple. I think the other thing is then go beyond demos. Yeah, really try to find the real problems and address them. Don't think about we do this tomorrow. You have to do this today otherwise it won't work. And then obviously you have to think in ecosystems. Yeah. Alibaba and also WeChat they have this uh extremely uh big advantage of being in ecosystems because Chinese companies always think in ecosystem we think in silos and these silos are not interconnected and that's a real problem but somehow we have to find ways to replicate this super approach um that they have that in our open web is uh uh bringing stuff in there with their uh protocols that maybe one point I don't know when uh we'll achieve this but that's also something that we have to work on and the last thing and I think the most important thing is optimize for trust make sure that people can actually trust your solutions yeah if it's if it's a demo that's nice yeah but I won't really use it yeah you can't scale users you can't really achieve the aim for growth in revenue if people don't trust it and very often we can't trust these systems nowadays. Yeah. And that's the reason why Boston Consulting and PWC and Mckenzie don't find this economic value. Very often the trust is lacking. That's something that we uh need to aim for.
And we have to realize that these guys that we've been talking about for like 2 three years. Yeah. Sheen and Temo that was just the start. It's not [snorts] stopping. Yeah. We see more waves coming in uh with totally different companies. Yeah. As I said, I work for automotive there. I already see the BYDongs and all these other Chinese companies are coming here with really great uh quality cars. And that's a very interesting thing or a very important thing to realize. Um Sheen and Teu are about let's say value. Yeah. Being cheap. Uh that's their competitive advantage. very often these new companies that coming in from China they will um bring in really good products and that will be quite difficult actually to um compete with them. There are also for example fashion companies that we will see from China that not bring in cheap plastic clothes but actually luxury products that will be very interesting and we always have to keep in mind when they are here they are these really digital organizations um that use all the stuff that we see with their no [ __ ] approach. Yeah. So stop getting surprised look to China uh for this kind of information. And I think and that's usually my last slide, not today, but usually um the best lesson that we can get from China is learn to learn. Yeah. Be willing to understand something fully before you know that it doesn't work. Yeah, that's especially in Germany, we are very quick in knowing why stuff doesn't work. Um but without fully understanding it. And uh as we just heard, we do another China panel tomorrow uh on the main stage next door at 2 p.m. um with uh three very able colleagues and we have a probably a very lively discussion and there's a little cliffhanger uh to get you there. We'll talk for example about the different modes of e-commerce that you see in China. Yeah, here very often we only see the searchdriven e-commerce and that will be handled by the AI agents. Nobody show up in your store when you only do searchdriven e-commerce. But they are alternatives and you find them in China. So if you can't make it tomorrow, come to our workshops. That's might be also an interesting alternative. You find this more on my website and you find the slides uh if you haven't copied or take pictures from every slide already. You find this uh on my um on my website under this uh URL. We have a very nice lively uh WhatsApp community discussing all these things. If you want to join uh you're welcome to do so. That's it what I wanted to tell you today.
>> Hi Bern. Um Anna from Avat systems battlesman. The question is I I saw these pictures with all these boxes and it's kind of a dystopian scenario. Who shall consume all of this?
>> Uh well um that that's a problem that nearly every company has right now. Yeah. And uh the thing is it gets cheaper when it comes from China. [snorts] So you can buy more. Uh yeah, of course sustainable is always an issue. Uh that's also something that I always uh um tell people when we talk about Sheen as a company. Um uh the interesting thing is that uh a western um fashion company often has the problem like producing 30 to 40% that never gets sold. >> Sheen is trying to not produce stuff that doesn't get sold. So that's more that's probably most sustainable is uh not producing anything that doesn't get sold. Um, but of course that doesn't solve the other problems. And one last question, do you see an opportunity when it comes to infrastructure, sovereign IT or cyber security from the European market? Because what we see there, I don't see us building these factories.
>> Well, the Chinese will build them if we don't. [laughter] >> And they actually already building them in Hungary, in Austria. Um, so it it's not like the Chinese only want to produce in China. They probably will also bring this uh here. that yes there are huge opportunities. For example, uh the Americans burn all this money on LLMs. Yeah. Large language models. Um, yeah, and there are people like Yan Lau, the ex research head of AI research at Meta who thinks that this is a dead end. Yeah. because uh we don't want to get too much into uh technology but it's all um uh uh it's not a deterministic but the probabilistic AI yeah it's all about probabilistic and so um they they will be certain limits that we meet where we don't get into can't realize the the applications that we want like for example robots probably won't work with these kind of uh humanoid robots with these kind of uh um large language models and They are trying to build world models. Yeah. That work totally different. Uh and that's a huge opportunity. Yeah. While the Americans are chasing the AGI using um uh LLMs, maybe the Europeans start working on totally different models, all the basic research that LLM is based on already came from Europe. So maybe the next AI will also do a little bit more than just uh providing um the knowledge. Yeah, that's I think one of the things that we can
>> Thank you very much. And uh these were mean questions so thank you very much for answering them. [snorts]
>> Do we have another question maybe over there? Um hi Bjorn thank you very much for sharing um well these ideas and a little bit like thinking of the future how the future is and I have a question of how do you envision we can create this bridge of building what's coming next in the sense that you mentioned that here in Europe for example we are silos and for me let's say a silo is this big company that of course they are worrying on surviving and of course on scaling and a silo can be also a startup that is trying to create the next innovation for the bigger companies but they are working in different paths right and in these super apps of course they are working inside the ecosystem and what I see one of the biggest challenges is that the startups are creating what they believe is the next step and the bigger corporations are worried in well growing surviving how can we bridge that so we can create these super apps because otherwise people continue always working in silos.
>> Yeah, we probably won't be able to produce the same kind of super app that the Chinese have like for example WeChat. Um uh I think Elon Musk is trying to achieve this with X. Um maybe OpenAI is actually uh trying to get there. Yeah. Putting all the stuff into the chatbt app. They actually want to create a a super app. But I think that's again a white wild white wild goose chase. Um more interesting is the approach that Google does. Yeah. With uh these open protocols that for example then handles the e-commerce thing across all the different players. That's probably an interesting thing. But uh before we talk about technology, I think it's more important to talk about philosophy. Yeah. that the um I think also a startup when they start they very quickly auto end up in organizing in in silos. Yeah. The people who are doing the sales, who are doing marketing, um who are doing product design, customer service, uh if it's just one person, yeah, it's not a silo. But if you hire a couple of people, you quickly put those people into departments and then you have silos and suddenly nobody is talking to each other anymore. And I think we have to overcome this. Yeah. The the the thing about Shien is not the technology. It's the fact that they or actually it is a technology but it's the fact that they have one consistent IT infrastructure that integrates the front end to the consumer. Yeah. The mobile app to the last element of the supply chain. And once you have integrated this you can build these kind of uh applications across all silos. um like Sheen Sheen isn't a super app and that's probably um stop thinking about silos is probably the first step to to to solve this problem.
>> Thank you very much. And maybe a follow up on that. I mean if you mention that of course the path to super apps it's not a viable path for Europe let's say but anyway uh Europe has to start stay competitive right so in their own ecosystem in their own world so how can these bigger corporations can transition into this streamline ecosystem that you're mentioning because I mean if they want to stay uh relevant for this future, they need to do something.
>> The problem is I only have one minute and I probably need a day uh to [laughter] address this. Uh it's a real problem, but right now we already have the problem of trying to think how can we uh get more independence from these American tech players. So we are already at a point where we think about change or we should think about changing things and maybe that's an opportunity to start getting to the bridge if we get to the other side is another question but uh that's probably a good opportunity to use