Michal Ukropec, CEO of twinzo, argues that there is no single digital twin but many parallel futures, from weekly drone captures of construction sites and real estate configurators to robot simulation and NVIDIA's Omniverse. Drawing on twinzo's projects in offices, factories, warehouses and ports, Michal shows why the strongest business case lies where things move, and how data connected to a 3D model opens the door to AI.
“there are thousands of potential tiny differences between what a digital, digital twin can be.”
“Because if you consider the typical manufacturing plant has a profit margin if of one to three percent, every single percent count, every minute counts.”
“because you have to have a good case, because it has to help the people. It should not be a persecution tool.”
Michal said there are around ten technologies, including cameras (only in plants and heavily regulated, not in office workspaces), Bluetooth, phone tracking, UWB tags and RFID.
According to Michal, it is a leadership issue: bring the unions to the table and make sure tracking helps people, because with a poor plant manager people will sabotage it.
Michal explained that a proper scan is evidence of how things were, showing oil leaks, water or mess, so many customers prefer nice 3D models and you should be careful what you promise.
Transcribed automatically (ElevenLabs Scribe) and lightly edited. It may contain small errors. Click a time to play the video from there.
Michal Ukropec: Ciao, ciao. I need to replace the picture. Sorry, five years old. Guys, don't you want to come to me? Please.
Michal Ukropec: Because sorry, but you know, there are just very, very few conferences where I like to do thing I usually don't do, and take a selfie. May I? Come, guys. What would I, what would I do without them? Thank you.
Host: It's also, it's also very expensive to take pictures with us.
Michal Ukropec: Uh, I didn't hear you. Sorry. Okay. Uh, do I have theee-
Host: Clicker
Michal Ukropec: the clicker? Actually, I have just one slide. Okay. Uh, really? I, I don't, I don't, don't need it.
Michal Ukropec: So, and they, uh, parallel futures. The futures word is new. Uh, I invented it today morning because, uh, you are all the surveyors and students and CEOs, and thank you. And, and, uh, sexy body. This was very cool.
Michal Ukropec: And very good student. I, I, I liked it a lot because I wasn't that good student, so because I worked, I didn't study. So the parallel futures, because, uh, Tomas showed us the things from or related to cultural heritage. Uh, Arkadiusz showed us the infrastructure, city level, and things like that, but I am coming from very, very different, uh, environment. And this is actually not NavVis, guys, percentage, okay?
Michal Ukropec: Uh, because usually when you need or when you want to digitize something, you need to have some kind of scanning. Doesn't matter. In this case, it's NavVis, but, uh, it's just a coincidence. It's, there, there is nothing particular about that. It can be Leica, Faro, whatever.
Michal Ukropec: The facility or whatever it is, is more or less static, okay? You take the pictures, you, you do the scan, uh, laser or, or, or Gaussian splat, doesn't matter right now. You have the documentation, okay? You update it after a year or two or something, and, and that's it. But what if the environment is, uh, dynamic, like a construction site?
Michal Ukropec: Do you really want to go every single week and, and ch- and check the scan and, and do another one? And, uh, is it not easier to, to have a drone which takes the pictures every single week, and then you have something like that, and you can compare, okay, this is last week, this is this week, this is two months ago, and, uh, and similar information? And it's much, much better than, than a scan because from, from above the, the things look really differently compared to the, um, human, um, height perspective. Or if you have the building already in place but you want to sell it, also this is a digital twin, okay? Uh, another Slovak guys, uh, luckily enough, but, uh, okay, I don't like the layout.
Michal Ukropec: I rather choose, uh, that one, and okay, and, and I want to change the colors of the bathroom and, and, you know, the, even the mirror is not, not as I want to have it, and so on. Also, these are digital twins, okay? So if someone tells you this is a digital twin and this is a definition, la, la, la, okay? Sorry, it's just bullshit. Uh, uh, really, really, there are thousands of potential tiny differences between what a digital, digital twin can be.
Michal Ukropec: I am coming from the industrial world, so, so I will switch to that t- topic, uh, right now, but even for some people, an Excel table is a digital twin, okay? So when you hear something like that, it, it, it can change your perspective. So, okay, and then, uh, you have the bill. And this is Visual Components. Uh, a simulation software, but it doesn't matter if it's Visual Components or Siemens or something else.
Michal Ukropec: We will sell, uh, we will see also something from NVIDIA, from the cell level to the plant level. You can program all the robots, so it's not just a simulation in the sense that what it should do, but it really programs the robots, so if there are five, so they don't collide, and what sh- should they do? Or if you have, in this case, again, a NavVis scan, these Ricoh guys, German tiny startup, and they can, uh, look at the, look at the scan. There's automatic segmentation, and they can, you can do the layouts in the software without, without touching anything in a very, very simple and easy way, okay? So, so these are digital twins.
Michal Ukropec: But then, uh, NVIDIA comes to you and says, uh, "Michal, this is a digital twin." Okay? This is the Omniverse because they want to sell the stakeholders the super cool, super nice pictures of the future site, okay? So this is from the, um, BMW plant in Spartanburg. This plant, at that time, wasn't built yet, okay?
Michal Ukropec: So these are not real pictures. You see no EVs, combustion engine, okay? Old stylish. Brum, brum. Okay?
Michal Ukropec: But for NVIDIA also, this is a digital twin. This is platform called Metropolis, where you track people, and there is a camera feed, and it checks the movement of the operator, so you don't bend your back too much, and you don't do, do this or that movement because, because it will harm your shoulders or neck or something in the future, and they want to avoid this. And this serves them, of course, as a material for, for future analysis. So okay, so you see the spaghetti charts. You will see more spaghetti charts.
Michal Ukropec: Do we have spaghetti for the lunch, Michal? Nah. Okay, so this is spaghetti chart, but also this is a digital twin, and again, uh, from NVIDIA, but it's funny way, so, so this is a real picture, but that one is not. Okay? And what do they do is they compare the movement of the real operator with the artificial one, and they can generate thousands of potential videos.
Michal Ukropec: Okay, this is what he can do. This is what he did. This is what he didn't. So they can analyze the movement of the operator and say, "Okay, this is fine. This is not fine," because this is serving as a quality check, okay?
Michal Ukropec: And you see go, no-go, it's good or bad, and then you can deploy this at large scale, okay? So I am coming from this environment. However, I want to show you also something sweet and cute from, from our past. And this is the very, very first Twinzo built in, uh, two thousand fourteen, Android four four KitKat for romantic souls. At that time, it was not even possible to take pictures from the screen because the Android didn't, didn't support it.
Michal Ukropec: This is a very, very first project because Twinzo at the beginning was about people tracking. We didn't know at that time there is something like digital twins. And you see again the spaghetti charts and moving people because we thought, okay, this could be an, a g-good case. Actually, it's not, okay? Uh, but then, uh, another guys came to us.
Michal Ukropec: This is, uh, one of the office building in, uh, Bratislava, and all these orange dots you see here are people, okay? And okay, this is Lukas, and Lukas is right now over there, and here you can see the temperatures and, and noise levels and things like that, again, in a digital twin. But then they come, "Michal, but we have, we have lo- little bit bigger building in Budapest." This is British Petroleum building, uh, obviously an office, so if you think that this can't be done in Europe, yes, it can. Okay?
Michal Ukropec: GDPR is, uh, just, uh, paperwork. It's not that you can't. Okay? And you see, okay, four guys at the bar, and, uh, they have plenty of meeting rooms and focus rooms, and, uh, they want to see the condition of those. So okay, Caspian, okay, it's free.
Michal Ukropec: Twenty-six degrees, ugh, a little bit too much, but okay, let's go there. It's for five guys. Now I'm seeking quiet, silent, dry place, so I can go there. And, and all this information are available to all the people, so they are ru-running on such, uh, such totems, so it's available for everyone. Okay?
Michal Ukropec: So it's, it's not just for one guy. And here you see again the environmental data for, for facility managers. However, also this is a case where we struggle a little with the money on the, on the business case calculation side. Uh, because this is in, uh, Abu Dhabi, the Golden Tower. The customer comes, "Okay, Michal, we don't know exactly where, where we will find the money, but we know we will find it.
Michal Ukropec: So we are going to integrate thirteen hundred sensors just in this one building." Uh, so you see who is sitting there, all the environmental data, everything, everything what you can just imagine that a smart building could, uh, could have on the level. And, um, yeah, you know, the Middle East is, um, in these things a little bit more advanced than, than Europe. You s- you see this not just on this or this kind of project, but also on the infrastructure. But then we see, mm-hmm, but the best business case actually is in the manufacturing/warehouses, uh, case.
Michal Ukropec: So this is one of the, uh, warehouses with the toys. It, it's very, very simple visualization from, uh, warehouse management systems. It's... These are very old pictures. Th-this is like twenty seventeen, twenty eighteen.
Michal Ukropec: Uh, so you see the, the difference between what we got at the beginning and what we have now, because this is something what you don't see in the plants usually, because they have s-still the paper boards, so-called FPS, fucking paper solution. Uh, it works like that. You-- So they take the pen, and they, they do the lines daily. Don't you think that your BMW wasn't done like that? Okay?
Michal Ukropec: This is very, very unusual, so they want to have all the data. However, how do you track something? Okay? And in this case, it's a camera, but it's camera useful for every case? Well, in this case, it was a topic, but in reality, you need to choose very, very wisely the technology if you want to track something, because most of your things are maybe static, but not in our world.
Michal Ukropec: In our world, the things are moving, and I want to see where the thing is right now, what did it do all day, how long did it take to, to, to get to A to B to C to D. Okay? And here you see thirty-eight information systems implemented because usually they do the integrations just in the MES, Microsoft Excel solution. Okay? This is the most common tool everywhere.
Michal Ukropec: So the Excel, Arkadiusz, is the single source of truth. Okay? Not, uh, whatever digital twin technology. And then they say, "But Michal, this is nice, but we have many plants." Okay?
Michal Ukropec: "We don't have just one. We have three hundred fifty." Imagine that. This long list of factories worldwide, this long. And we want to see one by one.
Michal Ukropec: We have smaller plants like ten, twenty thousand square meters to the bigger one, fifty, sixty, seventy. Even two million square meters. And they want to see how they perform, so they can compare and they, they can find the standard in the, in the, in the behavior of the plant. Because this world work like they're buying plants, they are selling plants. You know, like if you have something of three hundred in an Excel file, you are just moving the lines between, uh, here and there, and you don't care.
Michal Ukropec: So this is how the big w-world works. And this is actually the plant, uh, in the US we are going to visit next week in, uh, in Detroit. So this is, let's say, the, the typical level, but, um, when, when they want to have a really, really deep dive and they want to analyze the, in this case, logistics, uh, properly, using can see the heat maps, the, the spaghetti charts, okay? You can play the movie, uh, it, it like a movie you see, okay, this, this is the times of six in the morning, eight minutes, ten minutes, eleven, twelve, the guy still doesn't move. Okay, after, uuuh, almost twenty minutes, he began to work.
Michal Ukropec: Uh, Johnny, what did you do those twenty minutes? And you can begin to ask these unpleasant questions, and you can take it. This is how the manufacturing world works, okay? Uh, I'm always saying that automotive business is the fourth toughest after prostitution, drugs, uh, textile, and then is car manufacturing. You are laughing, but it's, it's, it's really hard.
Michal Ukropec: And this is Beko, ex-Whirlpool plant, now it, now it's Beko. But they want to see also the production, because if you have this production line is twenty years old and that one is five years old, and this is not having anything, you want to have one place where you integrate all the data points and, and see, okay, the production is running at ninety-nine point something percent. Okay, cool. This is what I have expected. Super satisfied, super happy.
Michal Ukropec: And this is how it was in time, uh, since the beginning of the shift. Okay, and this is post. Why? And the plant manager picks up the phone and again asks the unpleasant questions. Because if you consider the typical manufacturing plant has a profit margin if of one to three percent, every single percent count, every minute counts.
Michal Ukropec: This is Skoda plant, uh, in Mladá Boleslav, like one hour from here, uh, two and a half square kilometers. And here's, here you saw the energy consumption of the building. Okay, I want to cut the building into different floors, levels, whatever we, we call it, because I want to know, okay, right now we are producing Enyaq, and then the temperature is twenty-two degrees. Fine, I know Enyaq because it's my video. But, uh, I also want to know that the gate is open.
Michal Ukropec: Why? Because if you open the gate on one side of the plant and not the other one, in thirty seconds, your five hundred euros in heating is gone. Do you want to have every two, three, five, couple of hours, uh, throw the money out of the window? By the way, at Airbus, uh, in Toulouse, when they open the doors and they, and they are taking the parts, you mean-- I mean, wings and the bodies of the aircraft, it takes seven, seven hours to get the temperature back to the expected level. Or here they want to check the roofs.
Michal Ukropec: So they have the thermal scanners of the roof and, and they say, "Okay, and here we fixed this, this part of the roof at that time, and th-this is a, a highly risky thing. And here we found the water on the ground in, in August last year." Because they know that, okay, I found the water here, but it's not coming from there, it's coming from there. And they need to find the place. And this is okay.
Michal Ukropec: Uh, um, the, uh, the port in Amsterdam and again, uh, the, all the, the, all the shipping and, and logistics stuff, this is exactly the, a place where you see the big difference between the expectation and reality. You are, you are expecting the LinkedIn video with fully automated stuff. In reality, they are just throwing the things from one box to another. It's pure mess. And on the city level, of course, this is in Amsterdam, I want to see where is the ship right now and what the noise level and if the bridge is open or closed.
Michal Ukropec: Uh, the digital twin on the city level is always a topic. However, there is another problem, and that's the procurement process of, uh, of, uh, such a thing. If you-- and this is very funny, customers in the industrial, they don't want to see anything like this, because this is a, an evidence of how it was. But they don't want to show anyone how it was. They want to show how nice it is.
Michal Ukropec: Uh, that's why you see usually the nice 3D models and not your super proper scans, because they don't want to have the proper scan. The proper scan see-- shows, okay, here was an oil leakage, here was water, here was this mess, whatever. They don't want to see this. So when you will pitch this kind of tech-technology to your industrial customers, be careful what you are, um, promising them. Maybe don't want to buy such a thing, okay?
Michal Ukropec: Documentation is not always the best thing. Or they have a production plan. Uh, this is a very, very simple demo because we are now waiting for the data. This is brand new, like two weeks. And the thing is that they have a production plan.
Michal Ukropec: Actually, they have eighty thousand of production plans, and they don't know what it takes to produce one thing in the distance point of view. So, uh, they want to visualize it. This is very simple because they are not straight lines like, like you see here in the video, but they want to know in the end, okay, it's going to be six hundred fifty-two meters. And then when you have all these data points connected to the 3D model, you can begin to do the AI magic. Because then with, with the MCP model context, uh, protocol, if you didn't hear this yet, then MCP is something what you want to read about more.
Michal Ukropec: You can combine the stuff, the data points, convert it to a text, then talk about the 3D space with, with, with the LLMs, and this is the potential outcome. One of the possible digital twin futures. But which one be- which one will be the winner for you, I don't know. Thank you so much, and, uh, yeah, hopefully you saw something new.
Host: Michal, thank you so much.
Michal Ukropec: Yeah, sorry for the time.
Host: No, you're just perfectly on time. Uh, 34, 35 questions. You're really on fire, so let's just tackle them one by one.
Michal Ukropec: Sorry.
Host: So coming back to to the building.
Michal Ukropec: Yes. This is the most important question.
Host: Yes.
Michal Ukropec: And especially when you and your girlfriend goes there while you are married, okay? This is a real case, okay? But no one cares actually about the toilets. No one.
Host: No one.
Michal Ukropec: No, no.
Host: Okay. But are people tracked there?
Michal Ukropec: No.
Host: No.
Michal Ukropec: No, but, but, but this is very, very first question always.
Host: Mm-hmm. Mm-hmm. Um, all right. So you were, you were ready for this one.
Michal Ukropec: Sex first.
Host: Okay, so next, next one. How is information about a person's location actually collected? Do they wear some kind of trackers, or is it determined solely by see-through cameras?
Michal Ukropec: Uh, there are like 10, 10 different technologies you, you can use. Cameras, of course, as well. For example, uh, but cameras you can't use at the office workspace, just, just in the plant, and it is very, very regulated stuff. But you have different kinds of tags based on Bluetooth. You can track the phones.
Michal Ukropec: You can tr- track UWB tags, RFID. There are plenty of technologies. But we are always saying, "Please bring also the unions to the table," because you have to have a good case, because it has to help the people. It should not be a persecution tool.
Host: Mm-hmm.
Michal Ukropec: Doesn't work then.
Host: Absolutely. It basically, uh, leads me to the next question. How do you handle the ethical implications of data tracking in regards to people and their privacy?
Michal Ukropec: It's a, it's a leadership thing. If the, if the plant manager is an idiot, you see it immediately in the company culture, and this doesn't work. The people will sabotage it, definitely.
Host: Thank you so much-
Michal Ukropec: Of course
Host: Michal.
Michal Ukropec: You are... Thank you.
Host: Let's, let's give it up for him. Michal Ukropec.

CEO at twinzo | Digital Twins
CEO, serial entrepreneur, and passionate about the newest technology Last couple of years working on the “twinzo – digital twin,” bringing digital twins to everyone. To manufacturing plants, ports, offices, everywhere…