Navigating AI Integration

[00:00:00] Filip Popov: Hello, everyone, and thank you for joining us for another episode of Espresso 4.0. Today, I have a special guest, Jakub Kaczynski. Jakub, welcome.

[00:00:13] Jakub Kaczynski: Hello, everyone.

[00:00:15] Filip Popov: Excellent. Um, thank you for taking the time to take a brief coffee break with me. And, um, why don't you start us off by telling us a little bit about who you are and why we should listen to you?

[00:00:29] Jakub Kaczynski: Sure. So, as an industrial portfolio director, I'm responsible for having the answers that manufacturers need, that manufacturers expect around digitalization, industrial, industrial AI, all those areas that make a difference, um, in digital world helping manufacturers helping industrial customers to use and understand their data, of course, to increase their operational efficiency in various areas, quality, maintenance, production stuff.

And at the very beginning of my journey, I was, it was more technical related, technical oriented. I was working as an, uh, architect, uh, developer. Now I shifted to more blended technical and business related.

[00:01:29] Filip Popov: Got it. So, you are a blend of technical knowledge and business expertise as well. The perfect person to talk about, the perfect person to share their experience with us in the audience. Having said that, let's dive deeper, deeper, a little bit into that in your background. How did your journey from developer and architect to portfolio director shape your perspective on managing customer needs?

[00:01:52] Jakub Kaczynski: Uh, that was an interesting ride for me. Um, from the very beginning, it allowed me to be very close to the customers. And it occurred to me that it's extremely interesting to be very close to them, working with them directly, and learning the way that they work and they think about digitalization. And one of the biggest learnings, for example, that they had was being painfully pragmatic.

And now as portfolio director, my job is simple. I need to have those answers for manufacturers' problems. Uh, it's simple. It's not easy, not always, but you know, combined with this practical approach, um, learned during this, uh, technical part of my journey, it allows me to answer pragmatically, avoiding this, those buzzwords and hype, uh, and to connect what needs to be done with how this should be done.

[00:02:56] Filip Popov: Uh-huh. Fair enough. Yeah. Certainly, Industry 4.0 is, um, filled with buzzwords and trends and fads. So, it's very important to cut through that BS and, uh, find what pragmatically actually brings value and what it is, in fact, uh, the reason to onboard certain solutions or start certain projects in the field.

[00:03:25] Jakub Kaczynski: And we see that business is more and more tired with that, uh, you know, chaos in that space.

[00:03:32] Filip Popov: Yeah, it makes sense. You recently, now talking about your job. We actually had to move this episode and reschedule it because you were working on a project. So, recently, you mentioned that you're wrapping up a major energy monitoring and optimization initiative across 30 sites, I think. Could you walk us through the challenges and opportunities that came with managing such a large-scale project? I'm curious to know some practical kind of knowledge that we can share, uh, in part with the audience.

[00:04:02] Jakub Kaczynski: Sure. So, the main challenge in this particular case was that it's a capital group consisting of like 15 companies. And each of those companies working differently in a bit different industry. Um, so, it needs that the solution need to be flexible enough to address pains of on every stage of utilities, production, distribution and usage.

So, varying use cases, varying pains, challenges for every type of industry, for every area. And so, that was one of the main challenges to try to wire up all those, um, topics and outside of those, let's say, capabilities, when talking about large scale rollouts, it always boils down to looking at this from the perspective of customer value.

So, A, it's managing the stakeholders in such multi layered, you know, you have capital group, you have particular companies. Inside those companies, you have also multiple stakeholders involved, you know, from the top floor, shop floor, and so on, so on. So, finding every of those people have different goals, different needs, problems, and differently understand the value that this solution should bring.

So, managing those wishes is tricky. Balancing to meet also the second, uh, aspect. So, timely manner, right? Uh, because the times where, you know, manufacturers, industry can wait multiple years to see the result, those times have gone forever, I hope at least. Now, we need to be quick. We need to deliver value, you know, in weeks to months, not in years.

So, always such set up is very challenging. And if we wire up to that also change management, helping customers to prepare and introduce those changes, because that, you know, when you have that kind of overflowing solution in different, that helps in different areas, in different aspects.

It's a big organizational and cultural change to start working on digital.

[00:06:33] Filip Popov: Okay. So, I see the strategies that you employed and what were very important for you to kind of, uh, to run this project successfully. I'm curious though, from the aspect of the user or your end client, what do they need to have in terms of preparedness to fit like what would be an ideal case of preparedness from them or what should they do?

What do you expect them to do in order for you to do the things that you've just mentioned so that it puzzles together?

[00:07:07] Jakub Kaczynski: Um, in fact, it's, um, the answer to that is a bit, um, twofold. Because on one hand, uh, this particular customer did his job, he prepared the, you know, the overview, and he did some, let's say, homework, but we very often work with customers who just want to start, they don't know how, so their level of, let's say preparedness, as you mentioned, it's quite low, and that's fine, because, you know, first step is always the most important, but also the hardest one.

And our goal as a, you know, IT solution integrator as DPST is also to help those companies, to help them to prepare to that step and to perform it in the most efficient way, so not to waste money, not to waste time. Um, because when you start and you know, if you fail, that can discourage for years, sometimes.

[00:08:19] Filip Popov: Yeah, yeah, indeed. Okay. So, you would say that, uh, that starting something is the most important but also the most difficult part. To that end, you've emphasized that starting with customer use cases, rather than pushing specific technologies, is very important. Can you share a success story where this approach made a significant difference?

[00:08:43] Jakub Kaczynski: Yeah, of course, I have some of those, uh, lessons learned also, uh, some of them successful. Um, but, uh, basically it all, yes, it all boils down to having a real business benefit to the organization and the end users, not to have another shiny toy, right? Designed by the top floor for the shop floor. And, um, yeah, I remember one particularly important, uh, lesson learned for me, uh, because it was a very early lesson that also shaped my view on technology, those use cases, and customer value in those early days. Um, I was a young techie, passionate about industrial IoT because that was my entry point. And, um, at some point, we met a customer who wanted to implement some digital solution for monitoring the production, a little bit of anomaly detection, IoT, AI, and that kind of pretty usual stuff.

And they had all requirements gathered, they were ready, they picked the technology, and wanted to use Industrial IoT for data gathering, for AI, for analytics, and I remember that I was so excited, you know, it was the perfect project for me, that kind of requirements were ready, we are ready to jump happily into the implementation, and in the meantime, it occurred that this customer innovation team that we were talking with, they were also techies, very focused on technology.

So, you know, we had a great flow, great cooperation, but they had an internal, this team, this department had an internal policy that they always needed to have a couple of innovation projects running, like eight years back. So, so, um, a bit different approach still, right? Uh, and sometimes they were just pushing those initiatives to roll without this kind of, you know, preparing it without this building use cases, deeper analysis, cross-checking, challenging with, uh, with key users and so on that kind of stuff that now we consider obvious, right?

Just to have it running because they needed to have a couple of innovation projects and that was the goal. So, we developed this case, but without that kind of say appropriate context, it was useless. I mean, don't get me wrong. Um, it was doing its job. Yeah. It was just, it was gathering the data, visualizing some, um, simple analytics, but it didn't solve any particular important pain at all.

I mean, it was helping, right? It was helping with some minor problems. But none of them was the big one, the big pain that was, you know, implementing the solution was a relief.

[00:11:54] Filip Popov: Yeah. Yeah.

[00:11:57] Jakub Kaczynski: And how the project ended, we never leave the customer. So, when it occurred, we didn't say, okay, not our problem. No, we learned our lesson.

We did, you know, that kind of appropriate use case building. It occurred that it was a bit of luck as well, and there are specific use cases that might work. For example, it was around, if I remember correctly, unplanned downtimes. They couldn't find its root cause. It turned out that with minimal additional effort, the value was tremendous.

They managed to find the root cause of like 90 percent of missing reasons, right? Not so fun fact was that the most of those unplanned downtimes happened when operators switched CNC machines from the automatic to manual mode.

[00:12:57] Filip Popov: Yeah.

[00:12:58] Jakub Kaczynski: So...

That was that kind of, uh, final, uh, final of the story. Uh, but I remember this project until now.

[00:13:11] Filip Popov: What an uncomfortable discovery, but, um, but, uh, lessons learned indeed. It seems like the project was bound to fail in the end, but if you hadn't actually put effort into trying to identify what is important enough for them, in this case, a big, painful enough, if you will, you used the terminology pain point, um, it would have actually ended up in the trash. Instead, you've notified, you've, uh, discovered something that was painful enough financially, really, uh, where the application provided value because at the end of the day, we're talking about dollar signs or euro signs here.

[00:13:54] Jakub Kaczynski: Yeah, of course, everything has an impact on, you know, uh, on the financial, um, but it also requires that kind of problem that appears at some point, requires also some maturity on the other side, right? There are teams, I'm sure, I've met those that would say, no, trash.

[00:14:22] Filip Popov: Yeah,

[00:14:23] Jakub Kaczynski: Don't go further this way. Um, but usually there is something that can be done to, you know, mitigate that.

[00:14:34] Filip Popov: Indeed. Uh, going off a little bit on the technology and what's necessary for terms of architecture. In your view, what role do cloud and on-premise solutions play in manufacturing, and how can companies determine the right balance for their needs?

[00:14:53] Jakub Kaczynski: That's an interesting question, as I'm trying to be very pragmatic again. I believe that cloud services might have a tremendous impact. However, being trying to be real, I don't believe it will get widespread adoption in the industry anytime soon. I mean, that wide, really wide adoption, right? Uh, it's a more step-by-step approach, which is opportunistic.

So, looking for ways where plugging in cloud services can bring real value. And are more beneficial than doing something on-prem. So, for example, by using AI, right? Where training models can be more effectively done in the cloud, usually. Uh, so from my perspective, it's all again about those business cases.

Um, I know that I keep repeating myself, but, um, there are use cases where going hybrid is more valuable than staying on-prem. On the other hand, there is a lot of manufacturers that I, that still, uh, want to stay on prem, period.

Got it.

[00:16:16] Filip Popov: You mentioned actually AI now, and the use of, uh, cloud technology to actually get the power out of AI, right? To leverage it properly.

[00:16:24] Jakub Kaczynski: On a business case.

[00:16:26] Filip Popov: Indeed, going off of that, using that as a segue, what's your take on the role of AI in manufacturing today? Are there any common misconceptions you've encountered?

Yes. So, from my perspective, AI has one of the most important roles to play in the future of manufacturing. But again and again, it is a wise use of AI. So, it's not that AI is some magical tool that we plug in and, you know, plug in our operations, and it will make our dreams come true. Um, some time ago, we made an analysis of our past years of experience around manufacturing data.

It was like, you know, 200 projects. Projects in almost every industry for medium and large industrial companies, and we identified 30 reasons for the failure of data-related projects. Part of them was also AI, um, of course, it was, there was more, but we categorized drill down to around 30 categorized subjectively in seven areas

such as organizational problems, leadership problems, disconnection from business realities, and so on and so forth. And especially the last one, by the way, is from my, in my opinion, the grave mistake that is still being made. But we did that analysis to have baseline guidelines for our customers.

Um, and it also has shown how hard it is to roll, roll out, avoiding those pitfalls, right? And, um, and in general, there are a lot of misconceptions. One of them, as I mentioned, is disconnecting the data team from the business. So, building AI use cases without links to particular business problems that we want to solve.

And it still happens. Um, the success story, lessons learned, I mentioned, um, a moment ago, that could be easy, you know, AI, um, use case, AI project that was done without linking it. It's the same, you know, the same, uh, skeleton, the same base, and Matthew Littlefield, the president of LNS Research, it's one of the research companies that personally I admire most in the area of industrial transformation, Industry 4.0. Recently, like three days ago, he said on LinkedIn that 2025 will be the year of anti-hype. Now, breaking down the AI influencers, uh, hype, and so on, so on. Unfortunately, I don't believe it. I wish this had become true, but I don't believe it, unfortunately. Um, I believe a lot of hype will still, if we skip this, you know, in general, gen AI hype, we'll go into AI agents hype, that kind of event, unfortunately.

I believe AI Agent's hype is already here among us. It's in the room with us. Okay. So, I think what you're saying in terms of misconceptions if I can boil it down or distill it down to a point, is that people often see it as a glass ball, they can, they can like, uh, or, uh, or a magical lamp.

[00:20:22] Jakub Kaczynski: Magic.

[00:20:23] Filip Popov: That can give, exactly, that can give them answers at the click of a button, whereas in reality, uh, first and foremost, there needs to be sort of an infrastructure supporting it, like nice good data. And secondly, um, no less important. There needs to be a connection between the AI application and the data science team that's working on it and shop floor realities, or more importantly, business impact realities, right?

So, you need to know your business case.

[00:20:56] Jakub Kaczynski: Exactly.

[00:20:57] Filip Popov: Absolutely. Yeah. A hundred percent. Um, okay. And, uh, to conclude, kind of the more serious questions, where do you see the most exciting opportunities for digital transformation in manufacturing over the next five years, if you exclude all the hype?

[00:21:17] Jakub Kaczynski: Uh, somehow a bit contradicting myself, I still believe it's in AI. Uh, it has, you know, an enormous power to change the way industry works from the shop floor to the top floor. And appropriate use of AI will enable some of those trends that have emerged for some couple of last years like modular design, you know, micro-factories, mass customization, sustainability topics, automatic execution of operations, all that stuff that we are, you know, I would say dreaming of partially, uh, will be enabled by a wise use of AI.

[00:22:04] Filip Popov: Okay, excellent. I look forward to that. I can only speak for myself when I say that I am working tirelessly on bringing that future.

[00:22:14] Jakub Kaczynski: Yeah, that's, you know, uh, a role of also everyone who is sharing experience and educating and so on. On the other hand, what I, I'm thinking as a side note, more down to earth, because, you know, we can talk about new modular design, micro-factories, but again, being realistic, uh, you know, not going to happen anytime soon in the wide applicability, right?

So, being more down-to-earth, I believe that more and more factories and manufacturers will start to go digital. Um, because that is the need, right? And I know that we are when we are talking about exciting opportunities, we might, we mean that kind of mind-blowing, you know, the front page of the paper, newspapers, and so on revolutionary, uh, ideas, but looking from the other hand, you know, on the other hand, the reality is that for many manufacturers,

going digital, doing this first step we mentioned earlier is that mind-blowing innovation. This first step is the hardest, and it requires the biggest change in how the company operates and how it thinks about innovation. Later, it gets a bit easier, and for me, that is also, uh, an exciting opportunity that kind of, uh, still increasing the adoption of those ideas, right?

Uh-huh. Indeed.

[00:23:50] Filip Popov: All right. Well, uh, thank you, Jakub, for imparting your experience and knowledge to us. I would like to end these podcast episodes with two less serious questions to bring some levity and give the audience an opportunity to get to know you on a more personal level. Um, so I would like to start off by asking, outside of work, what technology or innovation excites you most?

[00:24:21] Jakub Kaczynski: That is hard. Outside of industrial context?

[00:24:26] Filip Popov: Sure. Yeah.

[00:24:27] Jakub Kaczynski: I saw a promo movie about a startup that is trying to digitize the scent. The smell. Analyze, distill, digitize, transfer, and then recreate at another location. And I don't know how real is that, but that sounds cool. And, you know, having this industrial background, I'm already thinking about how innovation can impact, uh, you know, manufacturing and production, for example, I don't know.

Stent-based quality control? Who knows?

[00:25:04] Filip Popov: Indeed. Indeed. I've seen the same. By the way, that was kind of trippy, but I don't know. I would like to see its application.

[00:25:12] Jakub Kaczynski: Yeah, still, you know, that kind of, uh, I wish that is true, but until I see that, I don't believe that it's so, so close. Indeed.

[00:25:26] Filip Popov: And, uh, my second question is if I called you at 2 PM on a Sunday, what would I be interrupting you in?

[00:25:36] Jakub Kaczynski: Most probably spending some time with my six-year-old daughter, you know, drawing unicorns and machines and assembling LEGO sets. That's a new thing or just cooking something tasty. Um, if not, there's a chance I've been reading some books recently. I'm reading a bit about ancient Roman history, for example.

[00:26:00] Filip Popov: Well, okay. Um, a lot of that to dissect there, uh, to be honest, but I'm most curious: What's your specialty when you cook?

[00:26:11] Jakub Kaczynski: Uh, that is, uh, that's hard, but I would say that uh, I have a couple, maybe not, not as a whole, but I have a couple, um, chef's meals, um, from Spanish cuisine using chorizo.

[00:26:29] Filip Popov: Okay, nice. And a little, a little paella.

[00:26:34] Jakub Kaczynski: Always good.

[00:26:36] Filip Popov: Always good. Excellent. Jakub, thank you very much for your time. Thank you very much for taking a break to have a coffee with me on our latest episode of Espresso 4.0.

[00:26:45] Jakub Kaczynski: Thank you for having me.

[00:26:46] Filip Popov: Hope we'll get a chance to do that sometime in the future. In the meantime, we'll find each other on LinkedIn, as always. Um, and for the rest of you, I hope you enjoyed the episode. Have a good one.

[00:27:03] Jakub Kaczynski: Thank you very much. Bye.