In this week’s episode, Erik is joined by Jan Burian, Associate Vice President and Head of IDC Manufacturing Insights EMEA. His core research coverage includes Industry 4.0, digital transformation, and IT and automation tools. At IDC, their focus lies in aiding IT experts, corporate leaders, and investors in making informed decisions regarding technology investments and business tactics.

Our conversation delves into the intricate framework of the industrial Metaverse. Discover how businesses are meticulously constructing this landscape, piece by piece, with Jan shedding light on various technologies and specific use cases. Moreover, we explore the fascinating realm of generative AI and its application in industrial contexts. Learn how it serves as an intuitive interface within the industrial Metaverse, effectively navigating the complex terrain of data saturation.

Give it a listen and enjoy the podcast!

Key Questions:

●      What is industry metaverse and how does it work?

●      Scaling and scalability of metaverse as digital twin of the enterprise world.

●      What do you see today in terms of adoption of generative AI and solutions that are used in the industrial environment?

If you're curious to know more about our Jan Burian, you can find him on:

Website: https://www.idc.com/

LinkedIn: https://www.linkedin.com/in/janburian/?originalSubdomain=cz

Erik: Jan, thanks so much for joining us on the podcast today.

Jan: Thank you. Hi, Erik. Thank you for having me.

Erik: Yeah, well, I'm really looking forward to this. I know that you've listened to some of our podcasts, so you know that we mostly are interviewing CXOs of younger tech companies or executives of more mature tech companies. But now I have the luxury of interviewing somebody who talks as an analyst every day. So I'm really looking forward to digging into a couple quite challenging topics with you. Thanks for joining.

Jan: I'm actually a huge fan of your podcast show, because it's like a huge source of not just information but inspiration for my work. So every time, I just listened to that. Not every time. But when I listened to that, I got so many ideas that I would like to put them on a paper and write a comment to that and to publish it somewhere. This is actually how I started my publishing work. Because I do write a lot of thought leadership or opinion pieces in which I publish everywhere around the world. Again, IoT podcasts for me are a huge source of inspiration. So thank you for that.

Erik: Cool. I'm glad to hear that. For me, it's likewise a great source of inspiration and an education just hosting the podcast. I think today will be no difference for me. So we're going to be discussing industrial metaverse, which I think is a fascinating topic. Because it's a topic, if you ask 10 people around the table what is industrial metaverse, you would get 10 different responses, right? And so just defining the tech stack alone is challenging, defining the use cases, figuring out how this fits into a business. So I'm really looking forward to getting into the topic with you. Let's start simple. So if you were to give a 60-second pitch of here's industrial metaverse, what would that pitch look like?

Jan: Alright. Right away. The way we define it in IDC — we're pretty much in line with the other tech companies — is that industrial metaverse is a highly immersive future environment that blends the physical and digital to enable shared sense presence, interaction, continuity across multiple domains: engineering, operations, supply chain. But you can actually understand it as a sort of digital twin, like a complex digital twin. But there's also that perspective of immersive environment, super or hyper realistic. This is something which makes it different from, say, classic digital twin. Then there's also the difference between the gaming on social metaverse, which is really totally it's a digital environment, right? But industrial metaverse is something which is really connected to the real world, leveraging the data from a real world.

We can also look at it as like a time machine. Either way, you can fast forward, or go forwards or backwards in time and try to analyze the situations. You can or the user can also, say, test the environment doing the simulation. So this is also very different from the classic metaverse. So you're right. Back to the first sentence, 10 people would understand industrial metaverse in 10 different ways. But that's the definition we try to stick to, really immersive environment, advance physical and digital.

Erik: So if I break that down into more discrete technologies, you could say that the industrial metaverse, it's built on a network of sensors that are collecting real-time data that are then integrated into some kind of platform that integrates those different data flows. Traditionally, you might just look at that as a digital twin through your laptop. So kind of a traditional digital twin application, but then the industrial metaverse would also include an enhanced front-end, an enhanced user experience where you're not just accessing this through your laptop. But as you walk through the day, in different ways, you are able to interact with that data landscape. Then the third element would be the AI element where, because you have this digital twin, you're able to predict activities going forward in time. You're able to understand why an activity happened in the past by processing data. Is that a reasonable way to think about the technologies, the sensor or platform enhanced user interface, which I guess could be a few things and then AI?

Jan: Yeah, I mean, when I think of the industrial metaverse, I'll maybe start with the platform first. Because it's not just like about the data coming from the machines or coming from the production assets. But it's also the outside environment where the data from a different software could come in. So from ERP, PLM, you name it. It's another thing. But also, there could also be the data coming from, I would say, outside the company. One of the biggest benefits of such a platform is the real contextualization of the data. So you can recontextualize it.

I would really start the platform first, visualization of what's on a platform and then the different, let's say, data streams, bringing the data into the platform. From, I would say, revisualization point of view, this could be also the AR VR technology. But at this point, it's probably not that super important because you can use just a real screen of your laptop. That's not a big thing, right? So maybe in the future, it's going to move somewhere into the sort of like a semi-virtual or, let's say, AR VR environment. But now, to really leverage the power of industrial metaverse even now, you don't need to have it.

Erik: Let me better understand. It sounds like this could, to some extent, be viewed as a rebranding of an existing technology, which is basically a data platform with some simulation capabilities, kind of digital twin capabilities that maybe existed 10 years ago. Of course, it has grown more sophisticated. But why do we need the term metaverse? Why not just say it's a digital twin, and we keep making incremental improvements on that so that you can access data in more ways and have more datasets integrated and so forth?

Jan: Because it's also about just like collaboration as well. You're right that its certain part could be more like, it's really like a virtual, immersive digital twin. The digital twins are without something for a quite time already. Then there's that, I would say, the element of the hyper realistic visualization. Something with Nvidia is providing mostly to these use cases like the BMW virtual factory. So without that, it's like a twin. My experience with the digital twin was always that it doesn't have to be a visually appealing thing. It could be just like — I don't know. It could be just like the data. It could be output of that, right?

Personally, I'm a person who was working or starting a career in factories. So that visualization always played quite an important, I would say, element in a way on the employee experience. The people were adopting the technology. This is what I also see, that the visual part is also very important. It wasn't that it was a necessary part of digital twin before. For me, industrial metaverse is really the platform digital twin enhanced with that visual perspective. The other thing is also that place for collaboration or for trainings. So I would say digital twin plus. Someone called it like digital twin on steroids, which is actually true. But that's the current status or the current situation. It's going to come in the future. We can maybe touch up on that a little bit later.

Erik: Yeah, well, let's get into then the users and the use cases, because I think that's where it then gets a little bit more concrete. I think that's a good point. If you think about a traditional digital twin, the user is going to be a fairly well-trained analyst of some sort or like an engineer that has a high degree of understanding of the system. It's going to have often a fairly small number of users who really understand the system and interact with it. Does that change significantly now? Are we talking about then the extending the touch points where you have a much wider range of people who are interacting on a regular basis with the system?

Jan: Yeah, I think that it is a real, like a virtual representative of the real world. Then almost like the number of users could equal the number of the people being part of the daily processes in a factory. Let's put it like that. So production planning, simulations, shop floor engineers, quality managers, let's say, maintenance stuff. Everybody could get their own, let's say, part inside that industrial metaverse. They can leverage data, contextualize the data. As I already mentioned before, it's sort of like a time machine. So they can also try to understand some consequences of something would happen on a shop floor.

It's practically sort of like a parallel world to their physical environment. Then, of course, with the number of use cases, it could be probably expanding. The more open this world to line of businesses are going to be. That's the future. This is just when you think about within the four walls. If you go outside the factory itself or company, then we can think supply chain, all the entire value chains. So there could be connections to your customers and also to the suppliers. There could be connection to the, let's say, third-party participants like service providers or the servicing companies. So I'd say I think that once you get out of this, let's say, really four walls, you can really exponentially expand for the number of use cases.

Erik: Yeah, well, that's interesting, actually. You mentioned maintenance. I was talking with an elevator company last week. One of the cases they were looking at was, basically, they have SAP. Then in Germany, they have a manager of a maintenance team. You can imagine elevator maintenance, it's a big part of their business. It can be fairly complicated when things shift. This guy in Germany, he logs on to their SAP integrated solution in the morning, and he adjusts schedules and so forth based on the data he sees. In China, that doesn't happen. In China, you have turnover every 18 months, so you don't properly train people. The guys have a high school degree. They don't use laptops. Then we were discussing, okay, why not have just a mobile app? Use AI behind it to do most of the scheduling, and then you're just telling this person, here's what the system suggests you do in anything that is not going to work here. So you have a lot of automation built into this.

So on the one hand, you are really greatly simplifying the solution. You're making it so that a maintenance engineer or a maintenance manager with a low degree of education can use this. On the other hand, you're putting the technology behind the scenes in terms of the AI and so forth so that they can actually do a fairly sophisticated job of scheduling. That sounds like it's starting to move. That would just be one touch point, right? But I guess if you multiply that by 500 use cases and you start looking at the data streams flowing into each other, is that start to build up into then the system that you would imagine as an enterprise that's properly connected or connecting the people on the ground and the people that are doing operations back with systems as far up the stack as SAP?

Jan: Yeah, I think this is a thing going to be a future. It's going to maybe take a while to get there and to also connect these, let's say, single use cases into sort of something like a bigger world. Let's put it like that. Because in my understanding of the future, this is really like a parallel world to our real world. So you walk on the street, and you get to the factory. You go through the doors, walk in, take a look at everything around like in real life. In the far future, you could be also doing this in the virtual world. You can see the environment in a real-time. Of course, there should be some sort of like a security barrier side or cybersecurity element that must be embedded in this. But that's how I think about the future. It's really like a parallel world. The services like elevators — there could be cost. It could be machines. It could be assets in the factories. It could be Oryx, whatever — one day, it's going to be really like a digital twin of the enterprise world. But this is going to take maybe a decade. So it's definitely nothing that is going to happen in a couple of years.

But generally speaking, what's also very important here is that scalability element. There are already some companies which already tried building the factories in a virtual environment first and used that industrial metaverse environment for simulations and, say, for scaling the factories in a rapid tempo around the world, especially environments like electric battery production, which is a very, very new thing. It's interesting on its own. It could also be very competitive. So it's very important to really scale fast. This is where this tool could help.

Erik: Yeah, if you look at the early movers here, I mean, you mentioned EV battery production. Is it process manufacturing? I guess that strikes me as the market that has the most sophisticated digital twins just as a foundation, because they're managing very heavy manufacturing processes where downtime is highly impactful. But would they be the big first adopters, or are there other groups that you look to when you want to see that—?

Jan: Yeah, I think that rather than process, these are more, let's say, like discrete manufacturers, mostly the ones with complicated engineering products. But typically, the earliest adopters of, I would, say most of these digital technologies is the automotive industry. Carmakers, car OEMs, this is where the most innovation are being tested. They are far ahead of the others than the electric battery business. But that also could be seen as a part of the automotive. From the process industries, that will be mostly the food and bev. We saw some sort of virtual factories or breweries, for example. Let's say, this is real-like environment which is really with, I would say, high-speed production. Also that, say, 100% control over the process. I mean, the digital twin of the process of the entire production plant is really bring those super benefits. Because every stop there is causing big issues with their supply chains or their value chains. These are probably the automotive and I'll say FMCG. But mostly, food and beverage. These are the earliest adopters.

Erik: Yeah, got you. So if we think about then what it takes to adopt, obviously, there's a lot of technology that needs to be adopted. So there's a certain challenge inherent there just in making decisions, in budgeting, in implementation and so forth. I guess there's also cultural challenges because you're then looking at maybe changing how a lot of business processes run. You're providing visibility into areas of the operations where there wasn't visibility before, which some people will react negatively to. What do you see as the main challenges, the main barriers to adoption?

Jan: I think it's pretty much the same if you look at the adoption of digital technologies in general. For many companies, if they're like a Brownfield, they have their technology working in the silos so they do have a data ready or data infrastructure, or even, generally speaking, the infrastructure ready for this type of projects. What's also interesting that I would say this is sort of like a paradox, some companies, on the other hand, they implemented so many digital tools or solutions, but they kept them running in silos. So the more digital they were, they had that feeling the footage of it already transformed but the more silo they created. At the end of the day, they were not able to really benefit fully from the power of the data, from the contextualization of the data. This is what I see as one of the biggest barriers: the way people are thinking about the data infrastructure itself, about how the data should be managed or should be processed.

Then next, another step is the way how to really turn the data into the insights. Because that what's counts, what's impacting the performance of the company the most. The ability of decision-making based on a data. Again, infrastructure, the way that people think about it, I mean, if you look at some IDC survey outputs, I'd say the sort of limited knowledge of what to do with technology is always among, I'd say, top three or five barriers. Of course, a lot of companies are also thinking, sort of not just afraid of being afraid. But one of their, I would say, sort of internal barrier in the adoption of this digital technology would be the concerns about the cybersecurity. Maybe some companies are even, let's say, they are — I understand that it's this element, especially right now where even manufacturing is becoming a critical infrastructure. But maybe sometimes it's this element which is really slowing down the adoption of some technology in a company, especially when we talk a lot about that IT/OT integration. Then while IT was, let's say, for a long time, it was the environment where the cybersecurity rules. The OT was left behind for some time. So now when it's everything online in general — you probably heard it so many times speaking to your guests on the podcast. But in reality, this is also one of the key barriers. So IT/OT integration, the way the data has been transferred, and also the security aspect part of this challenge.

Erik: Yeah, I know it's funny. Probably 40 years ago, if you were talking about a new technology, it would be the military or the Fortune 500 companies that were rolling it out the soonest. Now the challenge you have with this concept of industrial metaverse is, people think of meta. Basically, it's a consumer e-commerce. But the consumer just moves a lot faster in terms of adoption now. Although, I guess they're hitting some speed bumps.

Jan: Yeah, exactly. Because that was actually the reason I started writing the articles about industrial metaverse. Because when I visited some conference on some tech focus event, I had a feeling that it was, I would say, that this concept was misunderstood or there's misinterpretation of this term, of the entire concept. That was why I started thinking about it and then talking to the vendors, the leaders in this area. But also, it's not just about let's say like a software technology provider. It's also about the hyperscalers, to understand whether this is something they see as a technology which they believe in. Because at the end of the day, they are the ones who are really providing the infrastructure to the platform so you can have as many platforms. But when it comes to the data centers, there's not so many players on the market. So that was the reason why I write. As you said at the beginning, 10 different opinions, 10 different understandings.

Erik: There's one other topic here I wanted to touch on you, which is the topic of environmental sustainability as a cluster of use cases. I know you've written an article recently on the topic. Why is this industrial metaverse tech stack particularly useful for addressing the topic of environmental sustainability?

Jan: Yeah, it depends on the angle. But generally speaking, such a virtual environment helps you to, I would say, reduce the number of interactions or activities in a real world, which is at the end, a fact that saves. It's always reducing the carbon emissions. So you don't have to travel that much. You can share information. If you test something, you don't have to create the physical prototype. So you can do everything in a virtual environment even with just the new factory. So you can do or start simulations in a virtual environment using the industrial metaverse as also a collaboration platform.

So this would be probably the direct impact on that sustainability element. The next thing is, if you understand the industrial metaverse more in a broader concept, then, for example, in supply chain, it helps to get a better control over the supply chains. What's also important here is that, maybe there's going to be someone who'll say I'll sneak peek into the future. But this is not just about understanding the C02 footprint of your supply chain and getting this information through the metaverse into your systems. But in the end, the fact that it should be also the tool for which you can actively control or manage the carbon footprint or carbon emissions. Because one thing is just to have the reporting. The other is to really be able to manage that, to actively reduce it. So what are these are probably just a different discipline behind this. But this is where the industrial metaverse definitely could be a powerful tool.

Erik: Yeah, it seems like this drive to understand your own carbon emissions and also the emissions of your supply chain, it seems to be driving a lot of effort to connect your facilities and to connect data streams between companies, which you could say doesn't necessarily have to end up in being kind of a metaverse type of solution. But it certainly provides a lot more data, that previously companies were not really incentivized to collect such as, okay, we want to discreetly know each machine how much electricity it's using, or at least each production line so we can compare them. We want to know what our supply chains look like, and where materials were sourced, and how much energy our suppliers are using and so forth. That provides a lot of data into the system that can then also be repurposed for other analysis.

Jan: It could be also seen as, I would say, the environment where the closed loop information is being available and also an environment where the product could be tracked during its entire lifetime or lifecycle. So that's also very important. Because I would say the one thing is there's one perspective. That's reduction of C02 or carbon emissions. But on the other end, the circularity also plays a big role there. Because probably, the most efficient product is the one you don't have to produce in relation to the C02 emissions. So that's something which could be also — I would say that industrial metaverse is really like a complex digital twin of the entire environment. That's probably its biggest benefit.

You also said that at the beginning of this podcast. There are so many technologies even now which you can apply to the certain areas of manufacturer's life, let's say, or over the lifecycle of a product. This is so complex that you're not just able to control everything by separate tools. It's almost impossible, the complexities arising. So having the one-single environment, one environment with all the information, it might be also beneficial for users. This is also where the data is being stored and processed. But also, then it enables you to apply the artificial intelligence. Because you have to have a data in place and governance in place. So that's a perfect environment for leveraging the power of AI.

Erik: Yeah, okay. It makes sense. Our focus here is on industrial metaverse, but I also want to pick your brain quickly on generative AI since you're on the line here. Generative AI, I guess, listeners are going to be fairly familiar with it. But mostly, from either consumer application or maybe a white-collar office helping to draft documents or office work, or so forth. If we look at an industrial environment, then I think we start to encounter a lot of challenges in terms of getting access to proprietary data and so forth or issues with requiring more or less 100% reliability for the responses and so forth. What do you see today in terms of adoption of generative AI and solutions that are used in industrial environment?

Jan: If I abstract from the, let's say, the marketing, HR and these generic functions, then stick really to engineering, production, quality maintenance. So we see that even if the companies themselves are trying to, let's say, understand the technology and trying to think about the use cases, of course if they use some, say, ChatGPT, for example, this is not a proprietary technology for them. What we see, that number of IoT vendors, they're already starting. Some of them started to embed these models into their tools or solutions. Like in PLM, for example. So this technology could be seen more as a sort of — I mean, where I can see the biggest potential are these, let's say, copilots. Helping people or enhancing workers, engineers, private information, coming from that complex environment, for example. So copilots, definitely. By the way, copilots also could help to tackle the workforce shortage in some areas, in some positions. Because you don't need to have that super skilled worker, or super skilled engineer, or super trained engineer. You can have, let's say, sort of like a newcomer. I wrote an article about this. It was on a company and their service department. They were able to work also with the more junior engineers, where its enhanced information coming from this, let's say, copilots. So copilot solution, this definitely helped to improve or increase the productivity in certain areas.

The next thing would be — and this is also we see that some companies are already testing that, let's say, for writing the codes. That's something which is quite a straightforward thing. But also, we see that some technology vendors are trying to enhance their solutions in terms of creating autonomous systems. For example, on a robotic line, if some issues happens or there'd be some most likely issue, then the system could automatically change or reprogram the robot. This is where the generative AI could help significantly.

Next thing, and it's what I think is more generic, I think that these language models, they also democratize the knowledge. You can have an expert somewhere in Asia, and you have a factory in South America. There could be sort of like a language barrier. It looks obvious. But in real life — as I said, I started my career in factories — we had those problems all the time. Like having some issues in one plant, and we need to talk to some supplier or to the machine builder in the other part of the world. There was a significant language barrier, for example.

Now with these models, you can really have — maybe it's the next step. This could be all done autonomously. But now you can have also sort of like a real, intuitive, let's say, translator between those two people. So it's really something I would really call a really good democratization of the information. This could be also the part of the metaverse, or it could be part of the different tools. Either the people could interact now in more effective ways if they are not, let's say, being replaced by the technology itself. But then it's a different story. But if there's still like, say, a person-to-person interaction, this is where the generative AI could help significantly as well.

Erik: Yeah, certainly a person-to-person interaction. But this issue of data overload, if you really have a connected enterprise and you have a tremendous amount of information in there, any individual trying to figure out where's the right information, put it in the right format for me, make sure I'm not missing something, help me expand upon this if I need to be, that's a massive challenge, right? It's really hard to address that challenge with traditional user interfaces of scrolling through boxes and drop-downs, and so forth. You just have, more or less, infinite number of options once you get into these really deep systems.

And so if you can have an AI that can give it a request and it tracks down the right thing, and you tell it how you want to look at the data, and it presents it in the way that makes sense for you, I could see that as being a huge UX improvement just alone as a way to interface with metaverse systems. Jan, I think we covered a lot here on the metaverse. Anything we haven't touched on that is important for folks to understand?

Jan: Yeah, I think our conversation was touching the sort of, let's say, emerging technologies, something which is super new. It's constantly developing. For many people working in the daily operations, this might seem also very confusing. They might be having the feeling that it's too much. Because like industry 4.0, that was 12 years ago started, these conversations. During the times until today, there are so many new technologies coming to those people working in operations. I'm not talking about people in IT but people in production, quality management but also CIOs, also like CEOs. So everyone who actively works with technology, they have to deal with the terms like cloud, like artificial intelligence and so on. Those cycles are probably shorter and shorter.

Because we talk about industrial metaverse now. We talk about the generative AI. Maybe we're going to talk about quantum computing. So I think it's very important to find a way how we, as a people and also the people working in technologies, benefit from those technologies, understand how to benefit from those technologies, and how not to get lost in these, let's say, new tech, new terms. Because you can get to the point that you could just say alright, I'm just giving up. I can stick to this technology. This is very late that I would like to go to understand. Then the rest is something super futuristic. Maybe it might be good for me, but I'm overloaded. So this is the outside of thoughts I'm having on my head, on my mind, how to deal with that. I don't have actually the answer, but the point is that the organizations should be really having someone like — I really like that term of chief future officer, but more in terms of a person who is on personal environment, whatever, or someone who's really tracking those, let's say, new things and putting them together in context with, I would say, business and cited trends, and turns those into the, I would say, sort of like a snake of information, sort of like a normal people. So that's very important.

Erik: Yeah, I agree. Technology development cycles are getting more rapid. And we as apes are not getting more sophisticated, at least not at the same pace as our technology is. So we have to find new systems that allow us to cope with that complexity. Well, Jan, thanks so much for joining us on the podcast. I really enjoyed the talk with you today.

Jan: Thank you, Erik. Thank you for having me.