Building Industrial Software - and Companies | EP04

In this episode of The Connected Factory podcast, Rick Bullotta, a seasoned entrepreneur and advisor in the industrial software landscape, joins Alexander Krüger, Co-Founder of United Manufacturing Hub, to share insights from barely graduating college to founding and scaling multiple successful software companies.

Takeaways

Go-to-Market Strategy Matters:
A great product isn’t enough—effective marketing and sales are essential to success. Without these, even great products risk becoming "science experiments."
Extensibility is Key:
Building platforms that allow customization and integration makes software more flexible, enabling access to broader markets and addressing diverse customer needs.
The Role of Legacy Systems:
Industrial environments must integrate new solutions with existing systems. Respecting legacy systems is critical due to their mission-critical nature and high switching costs.
Convergence of Technologies:
Combining AI, IoT platforms, low-code environments, and simulations can simplify innovation and adoption, allowing non-specialists to create impactful applications.
The Power of Low-Code/No-Code:
Tools must enable quick and scalable application building for non-programmers, such as operators and process engineers.
AI in Industry:
Generative AI can assist in automation, predictive maintenance, and meta-sensing, creating opportunities for proactive, closed-loop systems.
Simulation-Driven Development:
Simulations rooted in physics, chemistry, and biology are key for designing and implementing new processes, especially in the absence of historical data.
Standardization vs. Scalability:
Industrial systems often lack the standardization found in IT environments, making scalability and cross-system communication more challenging.

Chapters

00:01 - 00:52 Introduction and Rick’s Journey
01:46 - 04:30 The Challenges of Legacy Systems in Manufacturing
04:02 - 05:43 Early Contributions to Industry 4.0
05:43 - 08:49 Building Extensible Software Platforms
09:00 - 10:53 The Convergence of AI, IoT, and Low-Code Platforms
13:40 - 14:49 Overcoming Adoption Barriers in Industrial Software
14:49 - 16:50 The Next-Gen Historian and Scalable Data Flows
26:00 - 28:50 The Future of Simulations and Automation
28:50 - 30:37 Closing Remarks and Key Opportunities

Transcript

Alexander Krüger:
So we are here today with Rick, one of the successful founders in the industrial software landscape. And I'm super, super excited to have you today here, So from barely graduating college, then founding two successful companies. Can you elaborate on that?

Rick:
That's it. You know, added that later, but I joke with people though. And this is legitimate that to be successful, I think in any pursuit, it's not just technical skills. It's not just great product and all that kinds of things. And perhaps it's the social skills that I've developed in university that have helped me a bit along the way. But yeah, you know, in retrospect, I probably wish I paid a little bit more attention in college, but nevertheless, I think a well-rounded set of skills is essential to be a startup founder, as I'm sure you're finding.

Alexander Krüger:
Did you major in engineering or IT? What was it?

Rick:
Yeah, so I started out as a mechanical engineer and I was sitting in a lab one day looking at a piece of broken metal under a microscope and said, this is not for me. I actually switched to, in a way more quantitative, was operations research and industrial engineering. So yeah, a little bit more mathematics-oriented, supply chain, some elements of that. That's where I ended up.

Alexander Krüger:
I fully agree. I think mechanical engineering also lost me when they said, okay, we have no piece of steel here. Let's now pull on it for five years. Let's see what happens. It was like, wow, this is quite boring. Irrelevant, don't get me wrong, but not for me. So I was also then going more production side of things.

Rick:
Ironically, I ended up in the steel industry. So I was cutting out pieces of steel to go be tested by mechanical engineers.

Alexander Krüger:
Yeah, actually. One of our first projects was also actually in the steel industry. Quite random because the industry is not seen for being innovative and high margin and digital, but a lot of learnings. What do you actually do in your spare time? If I see you on LinkedIn, it's also love biking and being outside. So you enjoy your well-deserved post-funding.

Rick:
If it has two wheels, I'd probably enjoy it. My wife and I are very much into mountain biking, skiing, snowboarding, things like that. So I try and take advantage of those opportunities.

Alexander Krüger:
It's nice to get outside. Perhaps I still need to get to that point in my life to be again outside and not in the office the whole day, perhaps.

Rick:
Well, in a way, I think I've kind of been able to get to a great spot in terms of through advisory work, I get to be involved with the technical fun, the business development fun, but I can get away, right? It's a more part-time engagement. I will say I tend to be a bit more active and hands-on as an advisor than your typical kind of advisory board roles. And it keeps me intellectually challenged.
I'm doing a lot of research stuff. I've got a little robot PLC and some other toys here, but lately a lot of stuff around AI and LLMs and, you know, basically constantly trying to learn.

Alexander Krüger:
Yeah, this is actually, I think, the key to don't stop learning and don't also like be true to what you really believe is. Engineers also tend to reinvent the wheel a lot of the times, like specifically if they try to program stuff. This is, I would say, a dangerous place to be in, specifically considering like the German councils and the foundations and like they think about, "We know everything better, but there's also like an IT world out there that you could also tap into and learn from them."

Rick:
I have a lot of experience in that world as well because I don't know if you know, in the mid-2000s, I used to work at SAP research. We were doing a lot of EU-funded projects. In fact, a lot of what we did at the Future Factory Initiative in Dresden was directly related to the founding of Industry 4.0. If you recall, our CEO, Dr. Cogerman, was integral to the whole original Industry 4.0 initiative.

Alexander Krüger:
Was this after you? So perhaps for the audience, you founded like an MES company, like super early on, then you sold it to SAP, right? And this was then after like the integration process?

Rick:
Right. So again, 2005 we were acquired by SAP. I was involved with, it was actually a little bit more than an MES. It was called SAP MII, Manufacturing Intelligence and Integration. So in a way, it was kind of a UNS, right? It was bringing together all these different sources, letting you build applications, composite views of them. And then we realized that if we have all this information aggregated together for people, we can share it with other systems. And that's what got SAP's attention. We could now populate other applications with real-time data. Then SAP Research, the team there was called the Future Factory Initiative, doing some really cool stuff ahead of our time around additive manufacturing and RFID, indoor positioning, and all kinds of cool applications.

Alexander Krüger:
Actually, back then it looks quite crazy, like all these things that are still new now or like the vast majority of industrial companies not yet implemented having that 20 years ago. Crazy! I'm not sure what I was doing back then, but this is actually crazy. Was then ThingWorx a little bit of a logical consequence? Because you're saying like the MII was like building applications and integrations, and then ThingWorx was like making it even more accessible?

Rick:
Yeah, two things. Actually, it kind of grew out of that whole real-world awareness. So my background obviously was very much in the factory setting, connecting utilities, some of that as well. And then this work that we did at SAP, it gave me the idea that the same challenges exist in other industry verticals. So how can we build platforms? Ultimately, every new initiative, the Internet of Things, whatever you want to call it—one of the key opportunities is to build the tooling so that other people can build domain-specific applications. And that's always been my background, whether it's from Wonderware to Lighthammer and ThingWorx—they were really all tools and platforms to let other people build applications. So the idea was, can we have a platform that would serve both the industrial IoT market inside the factory and the connected everything, Internet of Things? In retrospect, I probably would have done that a little bit differently, split them up a little bit more, but that was kind of the genesis of what we did at ThingWorx.

Alexander Krüger:
I think this is actually like the essence of building good software in manufacturing. There's always this upskilling aspect. You have these business problems in manufacturing, they're super unique, and only the engineer on-site level knows how to solve them. But he doesn't know how to program. He doesn't know how to build applications. The UNS—he doesn't know how to get data effectively and scalably. And I think providing them tooling is key.

Rick:
That's where I think we see that with any new technology. Think of the applications of AI and generative AI. Even for more traditional examples, like how many engineers come out of university now and know how to program PLCs? It's a very small number, right? There's an opportunity for a whole new way to configure these kinds of systems. My vision was always, you should be able to take it from the original design and simulation. You're designing a new facility or line or process or product. You should be able to take it directly from design to simulation to why not be able to hit a button that generates 80% of the application for you. Use those same simulators to train users, to train AI. There's this convergence happening now of all these different technologies. Pretty exciting time to be in the industry. But ultimately, to your point, we've got to find ways to democratize applying this, meaning you don't need to be a specialist in all 10 different technologies. When you call it low code, no code, high-level languages, platform—whatever those terms you want to use, it's got to get easier.

Alexander Krüger:
I also think this removing this specialness about manufacturing is also super important. Manufacturing tends to reinvent the wheel—from databases, programming, just like programming PLCs, which could also be on a microcontroller or like compiled code bases. And now we have this new stream of technologies—CI/CD pipelines, which already dominate the IT world. We have AI for code gen, but if we are still on the ladder logic way of doing things, there's not enough code base to actually get that running. So creating a common ground on technology is also what we like—SQL, MQTT, Kafka—IT but for OT applications.

Rick:
I think we could tie this back to your comment on who's doing these developments and what their skill set is. What a lot of people forget is that ladder logic was an evolution of hardwired relay logic. Electricians used to do the implementation of these control systems or were deeply involved in them. We shifted to a more digital model. Yes, we'd love to have less uniqueness in what we do in manufacturing, but the fact that these are truly mission-critical—if the system fails, people get hurt, the environment gets damaged, there's economic damage—that is a little bit different than a social media site or a pet food purchase.

Alexander Krüger:
Yeah. But this also super interesting because aligning needs to work is strange. If you ask a Kubernetes expert running Google, they'll say, "We never have 100% availability; we're cracking down on the fourth digit." But if you ask a SCADA engineer or factory IT guy, they'll say, "This cannot fail." It's not possible, but it's also how they think differently—failover, hardwired. If one application dies and cannot reach the other, the other application spawns. In a Kubernetes sense, they'd model it differently, which is super interesting.

Rick:
I think there's some mythology now with Kubernetes and similar technologies solving all those problems. The reality is the containers, the pods, the software running in them need to be designed to be highly available and scalable. The infrastructure doesn't magically make all that happen. It needs to be designed.

Alexander Krüger:
Yeah. The real-time aspect is key. For example, PLCs are designed to be compiled real-time and guarantee delivery of the code. This was one or two years ago when the main Linux kernel was patched with real-time capabilities. It was a good push during the supply chain crisis when Siemens PLCs were unavailable, forcing virtualization. But not everyone is there yet.

Rick:
Yeah, I think the initial effort around virtual PLCs is probably going to take much longer than we think. Now it's excellent for testing, right? Simulation and development—fantastic approach. But in production, for a variety of reasons, I think we're still a little ways away from that.

Alexander Krüger:
Yeah. There are also surrounding technologies that need to support real-time compute. Okay, we have it now, but real-time Ethernet, for example, also needs guarantees on that level. And there’s also a humongous change cost to do something like that. Everybody’s trained on Siemens, trained on the hardware, trained on the supply chain.

Rick:
It's interesting you say that because one observation I've had with a lot of the startups I work with is that they often talk about the challenges of getting their customers to scale their deployments—not just technical scale but rolling it out to multiple sites. And I’ve realized it’s not a matter of money. They can find the money. It’s the time.
Can the customer devote the resources? Can they retrain people on a new technology? These are really obstacles to overcome.

Alexander Krüger:
Yeah. This is also why software in manufacturing needs to be 10 times better to justify a change. A good example of this is the Pi system, OSIsoft Pi. They’re now moving toward Wonderware Historian in the cloud, etc. But—and I might not make friends with this—it’s just an overpriced database with no technical relevance to exist. But people are trained on it, trained on the APIs, and have spent millions on system integration, so they won’t change. You now need to be 10 times better.

Rick:
There’s another element to that. I hear a lot of discussion about the next-gen historian. What does that look like? The data store is just one part of what a traditional historian does, right? You still need that last meter of connectivity. You need to be able to talk to the hundreds of different things you have out in the field.

Secondly, you need that domain-specific user experience that allows a non-technical person or a process engineer to compare things, aggregate them, without writing queries and things like that. So, the next-gen historian will require more than just a good time-series database.

Alexander Krüger:
Yeah, it’s way more. But hopefully, the Unified Namespace (UNS) is an enabler for that. You solve the connectivity once, create data structures on an event basis, and then just get it in there. Then you can do means, averages, moving averages—whatever you need on the time-series database level.

Rick:
In a perfect world—and we’ve had some discussions online about this—I really don’t want to have to go to two different systems: one for the current state and one for the historical exhaust. But the idea is to create a single surface through which I can access all that information and manage security at one layer.

Alexander Krüger:
Uber broker.

Rick:
You can have one namespace. Sometimes the UNS or whatever is collecting the history already exists in some other system. That’s the unique thing about manufacturing. When you’re starting a new IT initiative, like an order management or inventory management system, typically you don’t deal with as much legacy.
In our world, legacy is real. You have equipment from OEM suppliers, lots of legacy systems, and it really changes how we implement systems. You can’t just ignore legacy and duplicate it, especially in environments like pharmaceuticals.

Alexander Krüger:
Not allowed to copy.

And how do you see the role of IoT platforms? When we started as system integrators before UMH, we implemented ThingWorx—an OEE dashboard, predictive maintenance, etc. How has that changed over the last 7–8 years?

Rick:
It hasn’t changed enough. That generation of products was unique and powerful. ThingWorx, for instance, had elements of SCADA, HMI, scripting, and event handling. But since then, there’s been a lot of innovation that future IoT platforms should leverage—AI, containerization, modern deployment methods, flexible data storage. There’s a lot of room for improvement.

Alexander Krüger:
Yeah, there’s still a need. For me, an IoT platform should focus on being a low-code/no-code application builder rather than persistence or data handling. Perhaps controversial, but Grafana could be seen as an IoT platform.

Rick:
I see two issues with Grafana. First, real value comes when things are actionable—human data collection, interaction, or rules-based execution. Grafana is excellent for visualization but not great for actionable workflows or cross-data-source drill downs.

Alexander Krüger:
Yeah. It’s great for tools in OT environments but was designed for IT environments. It’s powerful for monitoring applications, but it can’t be a good OEE board because you need drill-down capabilities.

Rick:
Exactly. And I think we’ve evolved even in AI domains. TwinThread, for example, is now closing the loop—running algorithms, optimizing settings, and writing them back to control systems.

Alexander Krüger:
What’s your favorite AI use case that’s currently in production?

Rick:
I’d say four areas: Generative AI for operator and developer copilots, meta-sensing (using vision or vibration for insights), and predictive/prescriptive closed-loop control.

Alexander Krüger:
Yeah, most corporates went down the large-language-model rabbit hole this year. Let’s see how that unfolds.

Rick:
Europe is more focused on high-quality software over marketing hype. If anything, the hardest part of building a company is the go-to-market aspect.

Alexander Krüger:
Yep, and it goes hand in hand with product development. At UMH, open-sourcing software has allowed us to ship to many users without a go-to-market loop for feedback.

Rick:
Openness is essential, whether it’s open-source or not. Products need rich APIs and extensibility to allow partners to add value and address a broader market.

Alexander Krüger:
Definitely. We’re nearing the end, but what do you think is the next wave in industrial tech?

Rick:
What frustrates me is focusing too much on individual technologies. The opportunity lies in rapidly composing applications from these pieces. AI will have a profound impact, but only if the platforms to simplify adoption emerge.

Alexander Krüger:
Yeah, especially with the labor shortage in automation, simulation upfront can narrow development efforts.

Rick:
I’d guess that 50–80% of industrial automation is cut-and-paste work. We need environments that boost productivity, not just for new applications but also for continuous improvement.

Alexander Krüger:
Definitely, definitely. So it was as expected, really, really nice to talk to you. This will be the last episode of the year. Everybody, have a great Christmas break and a happy New Year. Look also at the other episodes. This one will be a really good one, and there are many more existing and new ones that will also be released on the Connected Factory Podcast. Thank you, it was a pleasure.

Rick:
My pleasure. And have a wonderful holiday, everyone.