Implementing New Technologies - Part One - With Florian Beil

Welcome to the Trend Detection podcast, powered by Senseye, an industry leader in using AI to drive scalable and sustainable asset performance and reliability. This is a new publication designed to help you go away with ideas on how to achieve maintenance efficiencies.

 

For this 4-part series, I’m joined by Florian Beil, CEO at Axulus Reply, an accelerator to enable manufacturers to embrace the Industrial Internet of Things.

In the first episode of this series, we discussed the challenges enterprises face trying to integrate key technologies, whether manufacturers have fully embraced industry 4.0 and much more.

Please subscribe via your favorite podcast provider (Apple, Google and Spotify) if you'd like to be notified about future episodes and let us know your feedback by leaving us a review. 

Transcript

Key topics covered (click to jump to the section)

  1. Integrating key technology
  2. Has Industry 4.0 been fully embraced yet?
  3. The ambassador concept
  4. Where should you start on your digital transformation journey?
  5. Listen to the Trend Detection podcast

Niall Sullivan, Senseye: Okay. Welcome to the latest episode of the Trend Detection Podcast. I'm really pleased to invite Florian Beil on from Axulus Reply. We've got really exciting conversation plans all around Industry 4.0 and Axulus Reply's actually very innovative way of approaching this, which I've had lots of different conversations around this topic, but it's really going to be great to hear their perspective. But first of all, I'd like to invite Florian just to introduce himself briefly and then Axulus Reply, as well. That would be great.

Florian Beil, Axulus: Yeah. Super. Thank you very much, Niall, for the kind introduction and also invitation to this podcast. I'm really happy to be here. Looking forward to some great discussions. Now, my name is Florian. So I'm the CEO of a company called Axulus Reply. And what we're doing is we're basically offering an engineering system for scaling digital solutions in highly complex industrial setups like production, discrete or process manufacturing plants, but also chasing areas like logistics.

And I think the philosophy behind this is that we really want to make it as easy as possible for industrial customers to really reap the benefits of very concrete solutions without a lot of project work or team resources to be spent, but rather a real engineered, configured to site, scalable approach.

Niall Sullivan, Senseye: And I guess, first of all, what led to the Axulus Reply actually appearing or becoming there? What was the gap in the market which you're filling, I guess, as a first question?

Florian Beil, Axulus: A very direct answer, all this came out of personal experience. I was working for a large German industrial, also responsible for the IoT business there. And spending 10 years there working on new technologies like cloud, like AI, like AR, you know all the buzzwords, and speaking to many, many customers, I realized that, actually, all the customers we were speaking to were very open for digitalization.

Nobody would say today, in that field, digital is nothing for me or it's not important, but then you then ask the second question and so, okay, where do we stand on? What can we build on? The picture basically flips a little bit. So only a very few customers really were, at that point of time, able to really define a use case, use a certain solution, and scale it across their production network.

And this was not due to the lack of interest or importance, but it was really a process engineering problem they were not able to crack. And by the way, when I was working there and helping them, also was struggling to really drive the adoption because we were always in a technology ...

First of all, you had some technology, you had platforms, you have connectors, you can collect data and you can build applications, but actually, when you looked at it, what the core problem is is not so much about technology, it's rather about how do I translate all these technology building blocks into a meaningful use case for a press line operator or for a chocolate manufacturing plant end, and how do I translate it into something that he understands, where he sees the value add?

Then the second challenge you always had in this project is, even if I have now understood the use case, I understand the value add, but how do I configure now a technology stack which addresses that, number one, but then also configure that stack according to installation? Because the technological or system complexity out there in the field is so huge that, if I build a solution for plant A, this doesn't necessarily mean that it fits exactly like it is on the plant B because you have a different PLC, you have different maintenance processes, you have different whatever.

And that means, in consequence, you're always in this type of project approach. You send people, they do something, they code, they integrate, whatever, and this is not so much a scalable approach. The project is actually not what our customer wants. They want the product and it delivers a certain value, but we were never in that regime. And actually, that led us to think a lot about what is it what's missing, how can we support that problem?

And actually, when you think it through, it's a very complex problem, but it's a countable context problem. So it's there a finite number of PLCs out there in the world, there are a finite number of machines out there, and every complex complexity which arises out of this combination is, today, nobody would address as a project. So there are similar situations like, if you configure your car on BMW, you also get, potentially, a billion different kind of cars out of the system, but nobody would try to build a car as a project anymore.

So there are dedicated configurators and engineering systems which approach that problem and that's actually what I thought was missing, based on what I saw for 10 years working with customers. And that's actually how we then came to the point, "Okay, we need a system to support it. What does it look like?" And then actually, out of that, came Axulus. So that's how it came to live, actually, in the short sense, based on bad experience, which maybe might be, anyway, the best motivator for a new business model, but that's how we came up with it and that's how we built it, at the end of the day.

Niall Sullivan, Senseye: That makes complete sense. It's finding those pain points. I know you've identified lots of different pain points within there. An interesting one is actually the adoption of technology. So when you're first speaking to organizations, my guess would be there's a lot of different technology that they're using. Maybe they're not working together, there's not maybe a clear plan or maybe ... Is that the case? That'd be the case if we're talking big enterprises out there.

 

Florian Beil, Axulus: Right. Look, for sure, that's the case. I think most of the enterprises I work with, and especially the large ones which have huge, also complex organizations around the topic, so there are also many stakeholders involved, one of the first challenges they have is that, number one, there are a lot of parties out there.

So there are a lot of different plans. There's production IT, there's corporate it, there is strategy, there is innovation departments. And most of the times, they're not well coordinated around one common set of use cases. So which problems do we want to actually address?

The most of the times, I would say they're not also aligned behind the target architecture, target technology picture. So plant A does AWS, plant B, Azure, and the third one does, I don't know. So this complexity around the internal organization of the customers is also driving the overall challenge to drive and scale adoption of digital solutions.

And then, also, one other thing. So that's internal complexity, I would say. An external complexity, for sure, also for a lot of customers is that, within the field of digitalization and digital solutions, new technologies, there's a very high dynamic in the market, like, I don't know, hyperscalers which regularly update their microservices, innovative companies like Senseye, which bring up a very tailored, but high-end new technology into the market. And for a classical industrial company which produces cars or coffee machines or whatever, it's also very hard to keep up with that, what's happening in the field.

And I think this combination of internal stakeholder management, always type of project approach, and then the external complexity where most of the customers cannot afford a digitalization department which just keeps track of that makes it really hard for them to scale on digital solutions. And that's why, also, if we talk about exclusive, if I reflected in our product, that's how we also have this very clear role concept and different roles which do certain things, but in think and around one targeted library of use cases.

There's more than one customer that tells us, "Look, the benefits of this type of system engineering approach," I would call it, "is also that, as a central team, we are able now to steer the development into a certain target technology area." So we can define which building blocks are there. So there is Azure, not AWS, for example, now. And if you do that on PowerPoint, Excel, it's very hard. I think it really requires a system-based approach and that's what we're trying to do.

Niall Sullivan, Senseye: Oh, absolutely. You've mentioned then about how some departments don't have digital teams driving these forward. So this all leads into one of the questions here about Industry 4.0, which I've probably mentioned on every Trend Detection Podcast up to now at some point. But in your opinion, in your experience, as well, has Industry 4.0 been fully embraced yet?

 

And I'm coming at another angle, as well, where I was at a conference in Munich, actually, a few months ago and they were even joking about, "Oh, there'll be Industry 5.0 soon," and taking the nick a little bit out of that. But then later on in the sessions, they actually had a session around Industry 5.0, which I thought was slightly ironic. I don't think we're quite there, but have we actually fully ... Let's start from scratch here a little bit and say, has Industry 4.0 been fully embraced, in your opinion, or it starting to be embraced?

Florian Beil, Axulus: It's actually a very interesting question. For me, it starts with definition, what is Industry 4.0? Right?

Right, but it's my interpretation, what I try to bring across now, and maybe there are different ones, but for me, the term Industry 4.0, coined in Germany, I think tries to reflect the fact that, today, we have a lot of new technologies coming up, especially on the software and digital side, like cloud, like artificial intelligence, augmented reality, edge computing. There's a lot of dynamics coming up now and at the speed of a software development, which hits now industrial industry, which is actually tied to hardware manufacturing processes, so car manufacturers or electronic manufacturers and so on, so forth.

And I think what Industry 4.0 resembles for me is just the statement that, look, this new technologies, they have the potential to change how you do products. They can serve new use cases which have not been possible, technically, before. And the adoption of these use cases will reap a great benefit for industrials out there.

And in that definition, I think, to be very honest, Industry, 4.0 has been only partially adopted. Everybody, I think, understands the concept that there's a lot of potential, but these technologies, which I mentioned before, still, the hard part of it is maybe not even connected technologies or ramp up a tenant or whatever. The hard part, for me, is still translating these technologies into something which fits onto an existing production process and is understood by the operator, the plant head, which are all highly specialized production processes. It's a big difference if you make a car or if you make ice cream or if you operate an airport.

And I think that's the stretch we are in, that we have generic technologies like IoT, like cloud, whatever, which potentially can solve a lot of problems in these industries, but defining these problems and adopting the solutions, that's a whole different ballpark and I think that's where we're in. So I would say we're partially having adopted Industry 4.0, but I also understand that, of course, there must be regular new terms in order to keep the hype up, but the general tone of new technologies coming up being relevant for manufacturing for industry is still there and will sustain over the next decades, I would say. So that will be my long answer maybe to a short question, but I would say partially adopted, at best.

Niall Sullivan, Senseye: Yeah. I'll say this as a marketer. I think marketers probably have something to blame, at least partly to blame, for these new terms. It's like, "Oh, here's a new buzzword that's going to maybe catch on and catch attention," and Industry 4.0 has clearly worked. But yeah, no, I think that's common with a lot of people.

I speak to a lot of experts. It's still being embraced at the minute, so there's no point jumping ahead too far and thinking about other things, as well. One of the things I wanted to ... With that, I'm just thinking in terms of when you're speaking to people, is there resistance from people with these new technologies of not necessarily the blunt answer of, "Oh, it's going to take over my job," but more is there just resistance to change and is it a generational thing?

Florian Beil, Axulus: It's also a very good question. I think, for sure, there's some generational aspect to it. My generation or the generation of my kids, not to speak about them, they just grew up with all this or ... I mean, my generation is on the edge of ... I got my first computer, which was a disc computer, Siemens, back then I think Siemens Nixdorf, when I was 14 or 15. And if I compare that to my son, who is now 15, who got his high- end gaming PC two weeks ago, of course, you're talking about completely different worlds there. There was an internet when I grew up. The smartphone just came up when I was, I don't know, 20 or 20 something. My kids live with it now.

And I think there is, for sure, a generational aspect to it because the people are just more used to technology, but to be honest, I would not say that there's an inherent resistance to digitalization on the shop floor. I think where challenges arises from is that, of course, for sustaining a solution on the shop floor, it needs to be very concrete and very tangible, what problem it solves, what is the value add, what KPI improves, and how it does do it?

And that's, of course, also, and again, coming back to the IoT hype. So it's a difference if I speak about, I don't know, condition monitoring or predictive maintenance on a very generic level and then come down to a very concrete operational problem on a shop floor where they say, "Hey, my ..." I don't know, "I need to monitor my tanks," or, "I need to monitor my packaging machines. How do I do it? Which data do I take up? How do you represent it so that I can operate a better maintenance process, which, by the way, looks like this?"

So the complexity level between what is on the, I would say, marketing slides to the level you need to go to really have an understandable value add, that's the bridge you need to make. And if that's clear, to be honest, I haven't had any manager or operator tells me ... who understood the value and said, "I don't want to have it." But I think it's this translation work that is sometimes a challenge for a lot of companies, which needs to happen, but if that happens and if it happens at scale, I think that the adoption is there.

I saw a lot of development around terms like cloud, for example. When we started to talk to customers in 2011 about connecting machines to the cloud and data somewhere else, there was quite some resistance in the sense they did not know how to judge it or is it secure or not, but these things, I think, are gone now. We had Volkswagen three years ago announcing that they connect all their plants to AWS and will do a lot of apps there. So I think this has changed quite a lot. And now, it's really about adopting a technology into concrete value add and use cases.

And maybe, in maybe this last sentence on that regarding resistance and change management, of course, there's a change management involved because people which worked 20, 30 years in a production line suddenly ... or there's somebody claiming they can build you a software which basically can optimize your job. I think that's also a wrong interpretation, wrong approach to a topic.

What I would always do is, or what I think is a better way, is go to the shop floor, you try to explain what type of use cases can be addressed, and then you need to have a flexible configurable technology which allows you to involve the end user in a sense, "This is how we do it. This is how the UI looks like. Would you like to have it like this? How would you design the UI?" And it's not so much about that this changes the technology. It's more about bringing the solution really into the world of the operator.

So one example was in automotive, where we designed a solution. The line operator basically ... We started build the app, we collected the data, we did the AI, and we exposed the information in a, I would say, state of the art UI, and then the operator said, "Look, if I show this to my guys, they don't understand it because it's far too fancy, as we are used to, I don't know, 2000 HMI designs," so year 2000. So we basically down spec'd the UI so that it kept the same information, but looked ugly, in my opinion, but the adoption was higher because they had the same color tone, the same look and feel, and then the mental step for the operator was smaller.

This was, for me, an example where you say, "Look, if you don't talk to the operator before you deliver a solution, then you made something wrong." You need to talk to them, you to work with them to make it fit into their world, and then you can build on that and add more functionality, add more use cases, whatever. But I think it's not a resistance, it's just the way how you do it, which makes a difference. But I also think that this will be overcome, so it will be more natural to work with these type of things.

Niall Sullivan, Senseye: That's quite an amazing story, that you completely changed the UI just to fit around a particular user.

Yeah. That is really interesting. But one thing I took from that, as well, I completely agree that you need to involve the users. The last thing companies want is an enterprise level, senior level to say, "You must use this tool. You must use this tool, without any input from them." It's not going to be adopted, as you say.

Our approach at Senseye is to identify ... we call them PDM champions, I guess our champion users. And that sounds very similar to what you're talking about there, is to get more excited about the products internally and then they become your, let's say, advocate internally, as well, to other plants maybe, and try and close that divide, as you were ... Again, you mentioned earlier about a lot of stuff's done plant by plant, different systems, different PLCs, different cloud structures, and it can help maybe to bridge some of those gaps, maybe.

 

Florian Beil, Axulus: Right. Absolutely. I think this ambassador concept, as you described it, is for really important because, at the end of the day, you can call it digitalization or Industry 4.0 or 5.0, whatever. It all comes into life when somebody has a problem and it's solved by a technology. And this somebody needs to be able to speak with you. And I've seen too many big bang, top down digitalization projects, as I would call them, so a C-level person stating, "We are digital now," and sees one example and says, "I want to roll this out now through all my plans," and, "Please execute," and stuff.

It works in a B2C mobile phone environment maybe because I can put it on an Apple Store and then it can scale, but it's not that easy in an industrial environment. You need to configure foresight, you need to connect foresight. You need to also, I think, identify the production processes that need to change, based on a solution, because it doesn't help you if you have a great condition monitoring predictive maintenance application and you still run the same maintenance process. So you would do the same at the same cost, just with a cool UI behind.

So there's always inherently a configure to site adoption approach required. And I think managing this at scale is what makes a difference. So if you go to an end customer and say, "Hey, we have a solution," and it starts with the first brainstorming workshop, when you invite them, then you have mentally maybe lost already 80% of the customers because there was next project and next thing I ... next cost cutting project, whereas a templified, productized type of approach gives them a totally different experience to it. It's like, okay, they have a template, they already work, it's technically set up, they can configure.

So for us, or what I experienced also compared to my previous job, is the whole discussion with the customer changes quite a bit if you have these productized type of approach, although it's not a product, it's still an engineered product, versus let's talk about digitalization product. It's called this way. So I think there's a lot to reflect on, but I still believe that this will come, for sure. So there's no doubt that, in the future, digital solutions like Senseye can make a huge difference for the customers.

So I think it's not so much about technology, which is a pretty depressing message for people like you and me, which are very focused on technology and what it can do, it's more about the adoption and making it easy and seamless for the end user to use this technology in a way that he understands it, which is a different type of problem you're solving for. But at the end, this will all come together, I think.

Niall Sullivan, Senseye: No, that's really good to hear. We should be optimistic, really, and there's still so much to come, as well. You mentioned some of the augmented reality, something that's often brought up. Maybe that's going to be the start of Industry 5.0, but I don't want to be making predictions about things like that, anyway, from afar.

But what I wanted to comment on, just one more thing about Industry 4.0, maybe not the last thing. It could be the last thing about Industry 4.0, but with all these different technologies, how does industrial manufacturing customer get their head around having all these different technologies or what do they adopt first? What are their tips, let's say, on their digital transformation journey? Where do they start?

 

Florian Beil, Axulus: I think that's also a very important point, how you start with a customer. If a customer asks me, what I always tell them is, "Look, potentially, the technologies out there are huge." You can AI your whole plant if you want and have money in the budget and the time, but that's not the best way to start, in a sense, that you also need to.

First of all, the reality in industry and production requires you to be as much noninvasive as possible. So a customer cannot stop the plant for six months just to get digital and then start producing again, but you need to have an approach which allows a step by step, module by module, use case by use case, can adopt solutions which support the operations, and then you build on the data you gather out of that to build higher value add, more value add, more complex value.

And this digital transformation journey, for me, always assuming a customer is really at the beginning, would always start with, okay, first, let's have a look at your production process so it aligns, then identify the neural points or the critical path in your production and start with collecting data, connecting machines, sensors for the neural points, and you start to measure key cup operation KPIs, basically condition monitoring. But you do that in a way that, later on, you can use the data for higher-end solutions.

And so that means, the first thing, you go through, you define these machines, you connect them, you build dashboards or configure dashboards for the operator so they can see now these are the operation KPIs. Then maybe, as the next step, you add a rule engine where they can define, "Look, if these combination of KPIs is currently at the machine, then I have a problem, so please do something."

And then you go from there. Then you take them, "Okay, now, we have the data and maybe we connected already a vibration sensor or maybe we can feed it into a Senseye module," which then learns the pattern and, based on that, predicts a failure, not a rule which tells you, "Now, you have a problem," but predicts a failure, which then allows you, as an operator, to do even more than just condition monitoring, but you can better optimize your maintenance process.

And then, as you expand the whole network, then you can work across stations and say, "I'm not only optimizing your machine and optimizing your line down to AI-based and to end process improvements," that you train in an eye and say, "Hey, we know the good output and the bad output. We trained the AI based on that," and then based on the different safety end patterns on the machine, CAI gives you an indication, "Might now be in the range with that problem," and basically optimize the end-to-end process.

But this takes a long time. It's not like I install it and it's running. You need to take the people along the way. They need to understand, "Now, we have some data transparency. Okay, prediction, end-to-end optimization." Then maybe, right at the end, if it really works perfectly, then you have automation and say, "Hey, maybe I don't need even a person in between anymore, but I can run it by a system."

But that is, for me anyway, most of the time, for our future. It really starts with this initial small steps and that's what I would always recommend my customers and say, "Hey, start with condition monitoring, based on that start and also with predictive maintenance for individual stations." If you can convince the operators on that, then you have the playing field for adding more and more on the same scalable infrastructure, which is then, again, technology, but it starts with that small steps, I think.

Niall Sullivan, Senseye: Yeah, absolutely.

 

Subscribe to our Trend Detection podcast

We post weekly episodes on a wide range of different subjects including predictive maintenance, industry 4.0, digital transformation and much more.

Here are all the places you can access/discover the latest episodes of the Trend Detection podcast: