MVPro Media – The Vision Podcast #31
Guest – Dr. Wilhelm Klein, CEO & Co-Founder | Zetamotion
Dr. Wilhelm Klein, CEO and Co-Founder of Zetamotion, joins Josh Eastburn to explore how synthetic data, human expertise and emerging AI technologies are changing the way manufacturers approach visual quality inspection.
In this episode, Wilhelm shares how Zetamotion moved from motion capture into industrial machine vision and explains why complex, highly variable products can be particularly challenging for conventional inspection approaches. The conversation explores grounded synthetic data, the realities of deploying AI on the factory floor, human-in-the-loop inspection, the evolving role of quality control teams, manufacturing data security, and the potential for AI agents to make advanced machine vision more accessible to manufacturers.
On this page:
- Podcast player
- Sponsor information
- Guest information
- Useful links
- Episode chapters
- Episode transcript
Listen to the Episode
About our Guest
Dr. Wilhelm E. J. Klein is CEO and Co-Founder of Zetamotion, an AI quality inspection company focused on making advanced visual inspection more accessible and practical for manufacturers.
With a background spanning AI research, machine vision, sustainability and technology ethics, Wilhelm’s work focuses on applying artificial intelligence to real-world manufacturing challenges. At Zetamotion, he is helping develop AI-powered inspection technologies that use synthetic data to overcome challenges such as limited defect data, high product variability and lengthy deployment times.
Useful Links
Zetamotion:
- Website: zetamotion.com
- LinkedIn page: linkedin.com/company/zetamotion/
- Facebook page: facebook.com/zetamotionltd
- X profile: x.com/zetamotionltd
Wilhelm Klein:
LinkedIn profile: linkedin.com/in/wilhelm-e-j-klein/
Email: klein@zetamotion.com
Episode Chapters
Click onto the chapters to access the relevant sections of the transcript below.
1. From Motion Capture to Machine Vision
Start: 02:32 – “Well, it’s a classic founder story…”
Wilhelm shares how his background in AI and motion capture led to the founding of Zetamotion, and how an unexpected manufacturing challenge shifted the company’s focus towards machine vision and industrial quality inspection.
2. Synthetic Data in Practice
Start: 06:35 – “Yeah, the key item is really the synthetic data approach…”
The conversation explores why complex, highly variable products can be difficult to inspect using conventional approaches, how grounded synthetic data can reduce the need for large defect datasets, and what the process looks like when bringing new products into an AI inspection system.
3. AI Agents and the Factory Floor
Start: 13:40 – “What we are also working on… is to agentify some of this…”
Wilhelm introduces Zelia, Zetamotion’s vision for an AI agent capable of guiding manufacturers through inspection deployment, before discussing why the realities of the factory floor still demand real-world data, context and human input.
4. Human Expertise, AI and Ethics
Start: 17:34 – “Speaking of human factors in AI…”
Josh and Wilhelm discuss what increasingly capable AI inspection means for the people currently performing quality control, why Wilhelm sees the technology primarily as an augmentation tool, and the wider ethical questions manufacturers should consider as AI becomes more embedded in production.
5. Protecting Manufacturing Knowledge
Start: 21:26 – “Do you feel like manufacturers should be thinking more about protecting their intellectual property…”
The conversation examines how connected AI systems can capture valuable manufacturing knowledge, why data security and intellectual property are becoming increasingly important considerations, and what advances in generative AI and industrial AI agents could mean for the future of machine vision.
“Think of it less like a robot that waltzes into your factory and tells you, ‘you are obsolete now’ … but rather like handing your best people a supercharged magnifying glass.” — Dr. Wilhelm Klein
Episode Transcript
[00:00:00.050] – Dr. Wilhelm Klein (Zetamotion)
So you should be able to walk up like in Star Trek and say, this is my product. I want to inspect it for XYZ, and it will help you build the entire thing, including the hardware, including how to deploy, including the training, including the synthetic data, the trained models, the fully finished system, everything.
[00:00:20.490] – Josh Eastburn (host)
Welcome to the MVPro Podcast. My guest today is Dr. Wilhelm Klein, CEO and co-founder of Zetamotion. Wilhelm leads an AI quality inspection company developing practical machine vision systems for complex manufacturing environments. Combining expertise in AI ethics, sustainability, and industrial applications, he leads Zetamotion’s mission to make advanced visual inspection more accessible, reliable, and useful on the factory floor. His work focuses on human-centered AI, sustainable manufacturing, and helping manufacturers reduce waste, improve consistency, and build greater confidence in quality control. Zetamotion builds AI-powered quality inspection systems for manufacturers, The focus is on helping industrial teams deploy inspection faster and more reliably, especially in settings where defect data is limited and traditional approaches are too rigid or too slow to adapt. You can find them online at zetamotion.com. Wilhelm, welcome to the show.
[00:01:13.340] – Dr. Wilhelm Klein (Zetamotion)
Thank you very much, Josh. That was a beautiful introduction indeed.
[00:01:17.770] – Josh Eastburn (host)
Thanks for your interest in being on the show. I’m excited about the focus that Zetamotion is bringing to the show. As I mentioned, it’s been a while since we’ve touched on some of these topics. And I think given your background as well, you’re bringing an interesting perspective. Obviously, we’re looking at the technical aspects a lot of the time, not always, let’s say, the human or the cultural aspects and how the world is responding to AI in that way. So I’m interested to hear what you have to say about that too.
[00:01:43.160] – Dr. Wilhelm Klein (Zetamotion)
Indeed. There are some fascinating topics out there. Not always do they perfectly overlap, these kind of topics, broader human cultural philosophical questions, and then the very practical, down to earth, how do you actually make it work on the factory floor? But there are indeed quite often these questions that come up, especially in interaction with manufacturers, with those who are using the technologies. They do ask, especially when you are on the lunch break, right? Once you’ve already talked about all of the specs, okay, what resolution do we need? What is your defect size? What is the measurement tolerances? And then you go and have lunch and then that’s where the other questions come up. How much do you think in 10 years we will all still have jobs or what’s going on? You know, and that’s where some of that other part of my background comes in quite handy.
[00:02:32.030] – Josh Eastburn (host)
So I definitely want to get into that. I’m wondering in the first place, how did you end up working in machine vision?
[00:02:37.980] – Dr. Wilhelm Klein (Zetamotion)
Well, it’s a classic founder story and in some sense a classic deep tech story as well, because we started out of university with some cool ideas and a cool novel way of applying AI in the machine vision field. And in our case, it was motion capture. Me and the co-founder happened to do our PhDs at a department called the School of Creative Media, which is sort of like Asia’s answer to MIT Media Lab, which brings together people from all sorts of backgrounds. And the area that we saw an interesting opening was in motion capture, actually. So motion capture and 6 degrees of freedom tracking, where at the time there were 2 ways of doing it. One was classic motion capture, which is what you see from the movies, right? Where they wear the suits with the blobs and so on. And that’s triangulation-based. It’s extremely accurate, but it’s insanely expensive. You need so many cameras. That’s really difficult. And then at the same time, there’s also the extrapolation from just 2D vision basically, which was very messy at the time.
[00:03:46.890] – Josh Eastburn (host)
Mm-hmm.
[00:03:47.210] – Dr. Wilhelm Klein (Zetamotion)
Very inaccurate. certainly not suitable for, for anything where you need precision. And we came up with a way of using some of the bits from the triangulation approach, which is to reduce the noise in the real world and couple it with the, the capabilities of an AI approach. And that’s, that’s how we got started. And that’s also the legacy of our name. It’s still in there somewhere in Zetamotion.
[00:04:12.690] – Josh Eastburn (host)
Ah, okay.
[00:04:13.990] – Dr. Wilhelm Klein (Zetamotion)
That’s right. And yeah, so we, we had a good idea, right? We had, or we thought we had a good idea. We had a cool AI approach and we found, we, we had a good hammer and we found a nail to hit on with it. We also had some pretty good traction in the beginning of it, especially using this tech for pretty large-scale user interaction and crowd interaction settings in museums or festivals and so on. And that traction was all at the beginning of 2020. Yeah, 2020, fantastic time to get into anything. Launch any business. Yeah. Especially with large crowds, you know, that need— that went away rather quickly. But shortly after we came across a challenge that was put out by a company in the US, one of the largest roofing manufacturers, which was looking for a better way of doing computer vision or actually a way of making it work in the first place. And they’re making roof shingles, right? Asphalt roof shingles, which you find pretty much all over the place in the US. And their struggle was that their product was just way too unique, too complex, too organic. And they’ve tried different approaches and just nothing seemed to work.
[00:05:32.320] – Dr. Wilhelm Klein (Zetamotion)
And very serendipitously, we came across this challenge and saw that, hey, if we use the synthetic data approach that we had been using to apply it to this motion capture space, and we dropped the 3D tracking, but doubled down on the product or on the object recognition and analysis part, we can solve this problem, right? And that’s how we got started and that’s how we got into this whole space. So it was not the straight journey. It was not what you also sometimes find, right? The classic I’ve been working for this vision company for 10 years and I saw a clear need here and that’s why I spun out a new venture. No, we, the universe led us to this.
[00:06:16.430] – Josh Eastburn (host)
Yeah. Did it feel like a stroke of luck that you were approached by that company?
[00:06:20.160] – Dr. Wilhelm Klein (Zetamotion)
Yeah.
[00:06:20.810] – Josh Eastburn (host)
And so you’re coming into vision with a different kind of toolkit and where have you found that you’ve been able to make a niche for yourself? Right. There are lots of different types of vision solutions out there. Why has Zetamotion persisted since then?
[00:06:35.360] – Dr. Wilhelm Klein (Zetamotion)
Yeah, the key item is really the synthetic data approach and the business approach that we have been building on top of that. So what we see a lot out there currently is two models in the machine vision world. One is what we call the DIY kit model, which is usually some nice and very capable and expensive camera that comes with a certain software suite. Which you can buy and then build your own solutions. Usually if you fit neatly into some of the categories and some of the types of defects that the pre-trained models are good at, or the pipeline is already suited for, it can work quite well, but more often than not, you are not a standard case.
[00:07:20.760] – Josh Eastburn (host)
Hmm.
[00:07:21.080] – Dr. Wilhelm Klein (Zetamotion)
And then it’s, it can be quite difficult to have that responsibility of making it work. lie with the customer because it ends up in what we call research hell. You end up having to spend so much time capturing additional data, training more, fine-tuning until you arrive where you want to arrive. Or on the other hand, you can pay a lot of money for a complete bespoke solution that takes care of everything for you, but tends to be quite expensive. So The niche that we are trying to build for ourselves is sort of sitting in between providing what feels like a complete solution at closer to DIY pricing.
[00:08:02.320] – Josh Eastburn (host)
Mm-hmm.
[00:08:03.170] – Dr. Wilhelm Klein (Zetamotion)
And that’s enabled by our tech stack and also the setup that we have as an international team where a lot of our engineering sits overseas, allowing us to do a lot more workload for a better price tag essentially. So the areas that we particularly focus on, you already researched and brought up very nicely in your introduction, especially when there’s very little data available, when you have lots of variation, when it’s very organic, very unique. These are that we like to get into. Examples would be the shingles, for example. We currently see a lot of interest from fabrics. We have several projects in the fabric manufacturing world kicking off, and then also things that are traditionally very difficult, like shoes or apparel manufacturing.
[00:08:49.110] – Josh Eastburn (host)
Okay. So setting my assumptions aside, what do these different cases have in common that makes them good candidates for a synthetic data approach?
[00:08:59.610] – Dr. Wilhelm Klein (Zetamotion)
Yeah, primarily how data hungry they are and how much workload you would usually have to spend capturing a representative data sample. So take the shingles as an example, right? You have so many variations of shingles in terms of color coding, the distribution of the granules, the specific design and so on. Technically, if you want to achieve the levels of accuracy that these guys are looking for, you have to capture the entire defect catalog for every single type of shingle.
[00:09:34.790] – Josh Eastburn (host)
Hmm.
[00:09:35.610] – Dr. Wilhelm Klein (Zetamotion)
To also account for the amount of variation, because every shingle looks different, you have to capture not just 1 or 2 or 10, but hundreds, if not thousands of these.
[00:09:45.290] – Josh Eastburn (host)
Hmm.
[00:09:45.790] – Dr. Wilhelm Klein (Zetamotion)
If you use a traditional approach, right? And that’s where the synthetic data approach can really come in handy because what we use is what we call grounded synthetic data, which means we are not fully synthetic as in we are not starting with a completely rendered product. We are starting with ground truth, real images. The way I like to liken it is you can just imagine working with us is the same as if you’re working with a human that you are teaching to be a new quality control worker. Because a human, you don’t show the human 10,000 labeled images, right?
[00:10:21.740] – Josh Eastburn (host)
Tell them to memorize them.
[00:10:23.340] – Dr. Wilhelm Klein (Zetamotion)
That’s not how it works, right? You show a couple of representative samples and you keep going until the human says, ah, yeah, yeah, I got it. I understand now. And that means the human has understood enough of what you might call the semantics of your QC parameters to imagine, to be able to apply the same principle to another product that looks slightly different, right? And that’s essentially that synthetic data inside the human head, right? That’s essentially the same thing. And that’s what we do as well. So we take some representative samples as our ground data, and in that we ground our synthetic data and build from there, extrapolate from there, ask of course the user to verify that our understanding of the data is correct. that we understand the substrate as well as the QC parameters accurately. And very quickly we should arrive at a very close alignment between the understanding of the system and the operators. And then from there we can get started and can kick off a lot faster than you would usually be able to. We still take in more data and still fine-tune accuracy and capture additional outlier defects or other parameters that are occurring rarely, but it’s really the getting started that can be accelerated so much if you’re utilizing synthetic data.
[00:11:53.510] – Josh Eastburn (host)
Maybe you can describe that process a little bit for listeners who haven’t worked with synthetic data before. What is the interface to that system? How are they providing those parameters? Or is that something that they are describing to you in plain, You know, human language and then you’re inputting that into a system?
[00:12:12.080] – Dr. Wilhelm Klein (Zetamotion)
So for now, we are primarily working in a very close partnership model where we are trying to make it as easy as possible, which means we are primarily providing a human interface. So we are doing most of our onboarding pretty much like this on a call where they share the defect catalog and they explain it to us. They walk us through what it looks like. what the variations might look like, and then they share the raw images with us as well, right? That gives us the initial understanding of how to work with their materials.
[00:12:45.740] – Josh Eastburn (host)
Mm-hmm.
[00:12:46.270] – Dr. Wilhelm Klein (Zetamotion)
And then once the actual inspection station is in place, it works via an inventory page on our dashboard. So let’s say you are onboarding a new variation of a given product. You would just go to the inventory page, you would say, here’s my new product, let me run the onboarding scan.
[00:13:03.210] – Josh Eastburn (host)
Hmm.
[00:13:03.540] – Dr. Wilhelm Klein (Zetamotion)
And then they just put it in the station or make sure that it comes through on the production line and that they hit onboard and they have the ability to specify what is different in this product or what to look out for. And that’s pretty much it. From there we take over and then partially in an automated pipeline, partially it’s still very supervised. We onboard that new product, that new pattern or whatever it might be. And hopefully, that’s our promise, within 24 hours you can start inspecting that—
[00:13:37.040] – Josh Eastburn (host)
Wow. Okay.
[00:13:37.680] – Dr. Wilhelm Klein (Zetamotion)
Variation.
[00:13:39.120] – Josh Eastburn (host)
Fantastic.
[00:13:40.390] – Dr. Wilhelm Klein (Zetamotion)
What we are also working on, but that’s more of a future project, is to agentify some of this, right? We are piloting currently a tool that we call Zelia. Zelia is an assistant that, ultimately the goal is to have an agent that can live on-premises off the cloud and it can guide you from, I have nothing to here is how to build the inspection station. This is how to, what data I need from you and I will build the entire thing for you. So you should be able to walk up to it like in Star Trek and say, this is my product, I want to inspect it for XYZ and it will help you build the entire thing, including the hardware, including how to deploy, including the training, including— including the synthetic data, the trained models, the fully finished system, everything. That’s the goal. We’re not quite there yet, but we are on the way.
[00:14:38.500] – Josh Eastburn (host)
Very interesting. Who is that designed for? Who do you think is the target user for that?
[00:14:43.570] – Dr. Wilhelm Klein (Zetamotion)
Manufacturers primarily, and especially those for whom it would otherwise be uneconomical to do it themselves or hire the required people or to pay big solution providers. It’s aiming for those cases where it’s, it’s a matter of democratization, basically. So SMEs should be very, very excited about this product once it hits the market.
[00:15:09.080] – Josh Eastburn (host)
So smaller operators who are, let’s say you’re looking at greenfield projects then, right? Where they’re just getting started and looking for that kind of guidance on how to set up.
[00:15:19.330] – Dr. Wilhelm Klein (Zetamotion)
Indeed.
[00:15:19.810] – Josh Eastburn (host)
And that actually raises an interesting point that I was thinking about earlier while you were describing how synthetic data work. We’ve talked a lot about the line between AI systems and classical vision systems, or between software and hardware. And I wonder how you think about that. Certainly, you hear claims about what synthetic data is able to do. We can design the whole system just based on a CAD drawing, to maybe some more moderate perspectives, which are like, yes, we still need good quality cameras, and so on and so forth. I’m wondering how you think about that interface between the older or understood technologies and this more cutting-edge technology.
[00:15:58.350] – Dr. Wilhelm Klein (Zetamotion)
Yeah, I’m, I am clearly an optimist, so I think there’s a lot that is possible. But I see that a lot of the potential is still a bit further in the future. For now, the factory floor, I recently published an article on LinkedIn where I think the title was, the factory floor is where AI goes to grow up. Because this kind of, we can do it from a CAD model and you don’t need any data. That tends to be impressive in a demo, but if you try to make it work in a production setup, it can be very, very difficult to actually get there. There’s so many components to it that we have found being really important. It just requires a lot more additional input. We also always design our systems with a human-in-the-loop function. I do not believe that anyone can build a universal model that can just account for all of the variation that you see, all of the chaos, all of the context that happens on a production line so easily and hit a level of accuracy that truly elevates what you’re doing as a manufacturer. Now, if you’re fine with a system that performs at, let’s say, 80%, okay, that’s doable, right?
[00:17:11.720] – Dr. Wilhelm Klein (Zetamotion)
But usually that’s not the case. Usually In order to make it viable, you’re looking for 99% plus. And for that, you need all of the expertise, you need all of the context, you need real input data that you can build off of. I, while I’m optimistic and I think there’s a lot of very exciting developments, for now, let’s stay a bit grounded.
[00:17:34.660] – Josh Eastburn (host)
Speaking of human factors in AI, what becomes of the human inspector’s job In this picture? How do you see that role evolving as AI systems are able to make more and more of these decisions or recognize more and more types of defects in more nuanced situations like this? What does that role look like?
[00:17:55.630] – Dr. Wilhelm Klein (Zetamotion)
Yeah, that gets us closer to one of those lunch break questions where the question is, are you guys not putting a lot of people out of their jobs? There’s a lot of answers to that. One of the components or one of the answers is that In many cases, we actually see that inspection is not done by dedicated inspection staff, especially in SME settings, but people who have better things to do are rotated through and everyone has to spend 2 hours a week on doing the inspection and nobody wants to do it. Nobody likes to do it, but someone has to do it. Here we are just freeing up labor.
[00:18:32.700] – Josh Eastburn (host)
Hmm.
[00:18:33.120] – Dr. Wilhelm Klein (Zetamotion)
But Even in the cases where you have dedicated inspection, usually it’s more of an augmentation tool than a replacement tool. So the way I sometimes like to describe it is think of it less like a robot that waltzes into your factory and tells you, you are obsolete now, you can leave, I’ll— let me take over. I know everything better than what you do, but rather like handing your best people a supercharged magnifying glass, right? Or some spectacles that are drawing their attention in an automated fashion to leverage their expertise and make them so much more reliable, faster, and capture their true insights and their ability to connect what they see in terms of QC to the larger context of the manufacturing plant. So we see that mostly more of an augmentation tool. And in fact, we have not come across any case where it’s just a straight up replacement for human inspection. It’s always at worst, if you want to look at it from that perspective, you might be able to decrease your staff, but it’s usually more of an augmentation.
[00:19:48.930] – Josh Eastburn (host)
Yeah.
[00:19:49.120] – Dr. Wilhelm Klein (Zetamotion)
Sometimes it’s also, yeah, sometimes it’s, it’s, it’s, it’s of course also doing, making it possible for you to implement a solution where a human could never do it, right? Whenever you have a production speed that goes at meters per second, what human could ever do that, right? So there’s, the question doesn’t even come up there.
[00:20:10.870] – Josh Eastburn (host)
Are there these types of, let’s say, technology ethics questions that you feel like manufacturers should be asking themselves that they aren’t?
[00:20:23.510] – Dr. Wilhelm Klein (Zetamotion)
I think they are quite aware. Most are really quite aware and especially in recent years where that expertise has not been present previously, especially in larger companies, I see a lot of smart hiring for people who are very well-versed and they ask the right questions, especially on data security in terms of IP knowledge, in not wanting to have your knowledge become part of the the large next model, right? And just being able to apply whatever you have built for decades to whoever prompts the LLM in the right way. There’s a lot of good questions being asked. And on the wider philosophical or societal side, there’s also quite some awareness on data sovereignty or which approach one should or should not support or follow or how to apply. And I see a lot of good questions there indeed.
[00:21:26.380] – Josh Eastburn (host)
Do you feel like manufacturers should be thinking more about protecting their intellectual property from getting gobbled up by robots?
[00:21:34.200] – Dr. Wilhelm Klein (Zetamotion)
So in some sense you can think of it like that. Ultimately, what if you have an entity that has a lot of resources? and could build things, what is preventing them from just replicating your factory? It’s the knowledge, right? Surely whatever you’re producing, your unique competitive advantages are probably not in the actuators and in the rails and in the rolls and in the raw material that comes into your factory or how those machines that are sitting there are processing it. It’s probably in the fine details how to do it, how exactly to apply it, right? How to, what’s the right mixture? What’s the right temperature? What’s the right way of doing it? The knowledge, right?
[00:22:23.600] – Josh Eastburn (host)
Hmm.
[00:22:24.080] – Dr. Wilhelm Klein (Zetamotion)
And that knowledge to some degree has to be captured by us in order to do proper quality control, or it has to be captured by others who are trying to help you optimize your production systems, right? So especially the large IoT systems, there’s a lot of buzzwords that go around, but essentially any big efforts to capture as much knowledge as possible to create smart factories subjects you to that danger. Because if someone could literally take a snapshot of all of the collective knowledge that you have built, what’s stopping from someone else— using that knowledge and building a copy of your factory, right?
[00:23:06.270] – Josh Eastburn (host)
Yeah. That 10 years ago, that might’ve sounded like a ridiculous assertion, right? Insane. Yeah.
[00:23:10.770] – Dr. Wilhelm Klein (Zetamotion)
It’s not, not anymore.
[00:23:12.520] – Josh Eastburn (host)
Yeah. Yeah.
[00:23:13.620] – Dr. Wilhelm Klein (Zetamotion)
Even, even the connections or the, all of the things around it, your suppliers and your customers and so on, all of that additional context and knowledge, all of that can also be captured and replicated in some way, right? So this, I am very strongly convinced that this is an extremely important question for any manufacturer to ask. Where, what data is captured? How can you ensure the safety?
[00:23:37.770] – Josh Eastburn (host)
Where are we exposed? Yeah.
[00:23:39.450] – Dr. Wilhelm Klein (Zetamotion)
Most recently, I’m sure everyone is paying attention to the news, right? This is now what, the 20th breach where some LLM broke out of their confinement and just casually broke into some other company. This is a very important question to ask as well. If you are capturing my data, how do you ensure that someone else is not breaking into what you’re doing or into what we are doing using some of the latest technology out there.
[00:24:07.150] – Josh Eastburn (host)
Very interesting. Sobering, thought-provoking, but that’s why we’re doing this. So yeah, thank you for raising those really interesting points. Looking at AI more broadly, as I’m sure you do, are there any particular developments that you’re paying attention to as far as the future of machine vision?
[00:24:24.590] – Dr. Wilhelm Klein (Zetamotion)
Yeah, I mean, we’re still only beginning to fully utilize the capabilities of generative AI, especially in the synthetic data world. There’s many tools that you can utilize to synthesize the data. Generative AI is extremely promising here, but there’s still a lot more progress to be made. And the higher fidelity we can do for less compute, the better and the faster we can go because Ultimately, the principle still applies. The more data points you have, the better your models will be. So with better and more efficient gen AI and higher fidelity and less compute, you can create millions, if not billions of data points and train on that and achieve even higher accuracy even faster. So there’s a lot coming there. And that is on the generation side. But then of course, the entire agent revolution is fascinating as well. And there’s going to be a lot of amazing work coming out and we’re trying to contribute to that with Zelia. But there should be some amazing tools coming out over the course of the next few years. Yeah.
[00:25:31.490] – Josh Eastburn (host)
For Zetamotion, you mentioned Zelia as something that listeners should look out for. Where’s the best place to follow your work?
[00:25:40.090] – Dr. Wilhelm Klein (Zetamotion)
We are present on LinkedIn, of course, where you will be able to sign up to our newsletters or look at what kind of posts we have. We are also trying to be somewhat active on our blog, on our website. That’s of course the starting point, zetamotion.com. Those would be the two primary starting points. And then from there, branches to whatever additional information one might find about us.
[00:26:02.930] – Josh Eastburn (host)
Perfect. Well, I’ve enjoyed our time very much today. Anything that I didn’t mention that you’d like to bring up?
[00:26:09.180] – Dr. Wilhelm Klein (Zetamotion)
There’s so many topics that we could still talk about, and I’m very much enjoying the conversation, Josh. Sure, we could go on for many hours, but It’s been a real pleasure indeed.
[00:26:18.770] – Josh Eastburn (host)
Thank you. Okay, I’ll let you go. My thanks to Dr. Wilhelm Klein, CEO and co-founder of Zetamotion. You can find Zetamotion at zetamotion.com. That’s zetamotion.com. And on LinkedIn, Facebook, and X, where the handle is @zetamotionLtd. Wilhelm is on LinkedIn as Wilhelm E.J. Klein. Links in show notes as well. If you’ve tried to put AI inspection onto a production line, whether it went well or went sideways, we should really do a horror stories episode, shouldn’t we? Oh, that sounds fun. I would like to hear about it. You can reach me on LinkedIn or by email at josh.eastburn@mvpromedia.com. And if you aren’t already getting it, subscribe to the MVPro Media newsletter at mvpromedia.com/newsletter. This episode was produced by Big Robo. For MVPro Media, I’m Josh Eastburn. Be well.
