Implementing New Technologies - Part Three - 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 third episode of this series, we discussed the most in demand industrial IOT use cases, how manufacturers can scale these projects themselves 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. Examples of most in demand use cases 
  2. Solutions
  3. Defining scalability 
  4. Who are the key stakeholders involved in these projects 

Niall Sullivan, Senseye: Welcome to the Trend Detection podcast, powered by Senseye, an industry leader using AI to drive scalable and sustainable asset performance and reliability. For this third part series, I'm joined by Florian Beil, CEO Axles. Mix accelerator to enable manufacturers to embrace the industrial internet of things. In the third episode of this series, we discussed the most in demand industrial IOT use cases, how manufacturers can scale these projects themselves and much more. I hope you enjoy it.

And just to dive into those use cases, obviously, I mean predict maintenance is one, we're very close to incentivize up for obvious reasons, but can you provide some examples of some of the most in demand use cases, even for your good starting point?

Florian Beil, Axulus: I mean, as in the demand use case is actually an interesting question. I think at the moment... And of course, I need to give a little bit background. So when I talk about use cases... Now in this call, I talk about use cases or templates that we have as Axles. So we enrich the system with templates. I'm not talking about customer use cases. So for example, if tech builds a solution template, it's their template and it's them to talk about it. And these type of use cases for example, they build, are really specific to their wave handing lines for example. So it's a whole new discussion on that level when you talk to an electronic manufacturer. But from our libraries, I think what is... The use cases, which are striving are very industry specific but on a generic level, I think a lot of it starts, as I mentioned, with transparency and condition monitoring. So connecting critical assets, configuring dashboards, and also having a rule engine, which is a starting point for me most of the times.

Then there are, I would say, specific use cases on top of transparency which are higher end and related then to artificial intelligence based on the data. So of course, Senseye are definitely in predictive maintenance is still a topic all over the place. So audio, certain assets where we can predict failures and the technology you have there with Senseye is definitely a leading one, which will bring a lot of value add to the customers. But there are also now for example, a lot of things emerging around computer vision based AI. So for example, and I can speak openly about it, so we work a lot with the front for the airport, which at the moment... I mean, initially if you would've asked me two years ago, do you think in the airport would be a big customer of mine, given Corona, I would say no way.

But at the end, the opposite way around now, I mean as they have the hard time, they are very capacity constrained and they're looking now for digital solutions to be able to scale. And one of the cases there is computer vision base where you film with a camera, certain assets or the airports or the so-called dollies or the small wagons which carry your suitcase for example, from the ban to the aircraft and behind the computer vision there's artificial intelligence which classifies the dollies and tells the operators where are now, which type of dollies so that they can load or unload the next plane. So I would say this is the second class of use cases which are on demand, like taking the basic information you collect with IOT and condition monitoring and then add some artificial intelligence.

Might it be on predictive maintenance, might it be on computer vision, might it be on quality assurance and production, so auto detection of mechanical scraps of failures on products. And this is for me the next level I would say of use cases which are interesting at the moment. And then the next step for me where we also have quite some interesting discussions on templates, which are getting more complex to be honest, is then really applying this concept of data transparency, artificial intelligence, not only to one area, to one machine, but really to an end to end production process. So we call it AI based process improvement, which basically templifies a setup where you connect different type of machines, maybe along the line, different type of sensors, and you collect data and you correlate this operational data to target output data like quality or quality of the parts you produce.

And then you train the AI to correlate this input with achieved output and give the operator an indication, hey, you're now going into a combination of production parameters where... And production is across end to end process where your output might be in danger. So please readjust now. But these are more complex ones. But it really is I think then for me the next step of evolution and all the things that we talked about, the connectivity, clouding, scale computing, artificial intelligence on a station asset level, on the end to end processes, these are the topics we're actually talking about and where we also templify. And then of course we templify for different industries. So we apply these concepts to logistics, warehouse management to automotive, to airports to machine builders, which then is always a very specific representation of this template, but it's concepts are comparable.

Maybe the last element I would like to add in use cases, always the question. So there's a technology part which is one legal brick, one solution element in our templates can be an edge device where you train AI in the cloud and then you deploy it onto the edge really to have a short response time of the artificial intelligence output. But that's for me, maybe the same plus of use cases, just the technical implementation is a little bit different with a technology building block called Edge, addressing time critical issues on the shop floor. So this would be my answer to that. So I think it's very industry and very process specific, but on a very high level, maybe this would be the use cases we were talking about most of the time. So data transparency, condition monitoring, AI bay, AI on an asset station level like predictive maintenance like computer vision, and then AI end to end process improvement.

Niall Sullivan, Senseye: That's really interesting examples though the airport ones particularly interesting as well. Something you just, I guess, from someone from afar would never even contemplate. It has technology sort of driving behind it and how much tech... And complete with the issues that are being experienced at airports been... If anyone's been traveling in the last few months, they will know that all too well, including myself. But how much technology again, and I guess you're feeling positive about that as well, that that can really make the difference and really assist airports in this instance as well. And... Sorry

Florian Beil, Axulus: I think look, I mean it's also. I'm sorry.

Niall Sullivan, Senseye: No, you go. Sorry.

Florian Beil, Axulus: Go Ahead. 

I mean, I just wanted to add one thing because it's also an interesting thing I never would've thought before. Look, we started with Axulus as an engineering system for digital solutions in industry and it really also depends a bit on the maturity of the customer, how adaptive are they to this core idea? But if you're starting with digitalization, it's actually interesting but not there. They're not feeling the immediate pain because they didn't try to scale a solution yet. But what helps though is because when we talked to an airport, they looked at it and said, okay, we somehow believe you can scale digital because we don't have only Frankfurt, but we have 40 other airports in our network and whatever we build for Frankfurt, of course, we also want to scale somewhere, but giving you half an engineering system, can you engineer or can you configure a solution for us which does this?

And this was then the dolly topic. So there are also type of customers who say, okay, we believe you can productize it somehow, but I now need a specific solution for that problem. And that's where this whole template business came up for us where we doing these templates now, we publish them for the customer and they can expose it to all different type of installation sites. Which again, proves for me the point that a generic technology, or a platform technology as we call it, is interesting for the customer, but they really start to take when they see a concrete problem solve for them based on the technology. And that's how we came up with this airport topic because if you would ask me what is your problem at the airport, I would never come up with that idea. But as you know, the operators know that we can configure the templates very easily for them and they can see that, that's how it started.

So it also provides you a means of activating the innovation of the real operators and the experts that they see, yeah, they can handle the technology, they can translate it into something useful for me. So what I need is this. And then you start to build a template, you configure it, you build it and you scale it. And that's, I think, also a very valuable approach to digitalization after all.

Niall Sullivan, Senseye: And I guess what I wanted to add, I mean it obviously will depend on projects, so it's not really a general view because every project is different, every template is different, every use case in some way... Well, maybe not different, but I guess what my point is, how difficult is it actually in reality to design some of these solutions end to end, as you put it? And I know that that's a broad question, considering the different projects involved, but maybe just a... Yeah, I don't know if there's an example you can give on that.

Florian Beil, Axulus: So I mean, so really taking an existing template and configuring it for a specific installation, if you have all the components I would say already in the basket. I mean, that's rather a simple approach for us because the whole end to end system excellence would guide you through that. I don't say in... I wouldn't say in minutes, but I could show you now in half an hour. Okay, we take the template, we configure it, we get the installation workflow, we publish it to the end user and they start installation. But of course, I need to say there's an assumption behind it. So the assumption is, okay, we already have a template, we have the technology build up and we can work with that. But if you start a new use case, we need to build a new template, maybe you need to develop some of the software.

So for example, the computer vision, part of the proper case I showed you, needs to be built once. But the big difference to my previous life actually was that with Axles, if you build it once, you have to legal break into your repository. That means next time you need something like [inaudible 00:12:25] injection, I already have a pre-configured AI model which I can reuse and reconnect into another use case. So for example, if another airport addresses us, but an automotive OEM who needs to deliver or which needs to deliver his auto components into big boxes, they have the same problem. They ship it all over the world. They have some areas where they park them but they don't know exactly how many are there, which type and so on and might ask, so can you adapt this solution to our problem, maybe not dollies but shipping boxes? And that means we can reuse, to a large extent, this model approach.

We can reuse a lot of the components of the previous solution and we can only delta develop the one thing or the UI or the training of the AI model which is missing. And we do that only once and then we have another legal brick in our repository which we can reuse. That means although you might at the beginning have some initial effort to build a template, the more templates you build, you get faster and faster because you start to build your modular repository, your integrations and your installation workflow libraries so that the next time you're much faster and you get faster and faster. And I think this is also a very interesting scaling effect, which the first customer start to see now that it's not only a good procedure and transparency in it can analyze, but you also start to build a repository of software integrations knowledge, which allows you to get faster and faster with each and every installation, each and every use case.

So asset, if it's there, it's very quick. If I build a solution, it can take, based on the UI components, I think the solution we build within or template we build technically not only Axulus but also with software components within three months. And from there on it starts to scale because then you have everything and you can configure it for different installations and you have to... Especially also in installation workflows in the system. So that's the timeframes we're talking about. But you will also get faster and faster. So the three months, maybe in one year down the road or two years might be one month or even faster because we have already a lot of libraries and bricks or our customer has that. So we also expect that you have quite some efficiency gains there.


Niall Sullivan, Senseye: And it is the idea with the scalability, I'm just thinking... First of all from Senseye perspective. So we have a customer on the [inaudible 00:15:01] essentially at a stage where they're so sort of advanced and mature, let's say, that they almost through... We have a platform called Omni-verse, which is provide step by step guide to scaling and they almost roll out new production lines themselves. They've almost taken on themselves. Is that the idea that you create the templates then when they get to the point of scalability, that's when you take a step back and they can reuse the templates themselves? Or is your involvement still in there in the scalability part of the phase, that phase?

Florian Beil, Axulus: I mean, to be honest... Yeah. So to be honest, I mean the idea is really that we provide the customer system where it can use it by himself to scale a solution. So what we typically... If you talk about an excellence core customer, so a customer who buys the engineering system and starts to build his own libraries, what we are actually doing in the onboarding is we enable him to use the system to build own templates, to connect his own repositories and libraries so that he can start to build and deliver his own solutions. That means we are actually not involved anymore in an execution of an installation workflow when architect delivers a solution. We are just now enabling them, initially, to use the system, to build up the libraries and to work with it. But these are the type of customers that are, I would say, mature enough to understand digital solutions, to build the libraries and to deliver them.

There are other type of customers maybe, [inaudible 00:16:40] at the moment still is where saying, hey look, I understand you can scale solutions, but actually I don't want to build a library yet or I want to work with the system but I'm not coding it but we would need a support from someone who builds a template in excellence for us so that we can later scale it out. And that means that you always have a partner involved that builds a solution, makes it ready for scale out in that scenario. But the general idea is really to help our customers to be independent and scale with a system by himself. And then it's also not our interest to... I mean it's their IP at the end of the day, what type of solutions, how they're built, how they're configured. So we're not... Unless the customer wants us to be involved, we are actually out of that and supporting just the core engineering system.

Niall Sullivan, Senseye: Yeah, and bear in mind the current situation. Well, yeah, obviously across Europe and the UK as well with energy prices, sustainability, and obviously with things like climate change, it's always been a focus, but I guess when it's really impacting and hitting home, I guess that's really starting to make an impact on the manufacturing industrial world.

Florian Beil, Axulus: Absolutely. 

Niall Sullivan, Senseye: Makes sense. 

Florian Beil, Axulus: I mean, I just kind of align that, Axulus is also in that sense, quite an interesting way of sharing knowledge across organizations, across people, and which also drives a lot of the success of excellence because at the end of the day, if there's a lot of interest by other parties looking or by different plants looking into the system and saying, hey, what's in there? Which could help me? What type of solution installation workflows the guys of the digital leap plot that do and can I replicate it? For me, that's also as a motivation to use the system and preserve and share the knowledge that the organization has. I fully agree.

Niall Sullivan, Senseye: And digging in some of these projects. So who are the sort of key stakeholders involved? And again, it depends on the project, but on a very high level, different areas of the business, who needs to be kept informed and be apart from the very beginning, let's say?No, that was a really, really great.

Florian Beil, Axulus: Right. Look, I think on a very high level, I would always maybe distinct three roles. Of course, one of the key roles is the end user, or as we call it, solution consumer. So the head of... Production heads, the plant heads, the line operators, production IT guys, which are really looking for digital solutions to address specific problems in their environment. And they are the ones which are looking for use cases which configure their requirements and then at the end of the day also supervise at least the execution of the installation in their plants. Then the second role I think are the experts that can configure solution stacks or then also maybe forward a required module for development into an R&D process like coding. So these are then really the... I would say the classical IOT developers or software development teams which would do that.

And the third role, which is important are the ones which are executing then the installation on site and enablement, which can be the same then, the solution consumers or the end user. But sometimes you have also external service providers who do the hardware installation or the enablement.

So these are the three parties which need to come together in order to bring a solution to life and assets. Typically, these are plant heads, line heads, head of production systems in the first place. Second part, would be central digitalization teams, development teams if in the company or external partners who do that. And the third parties would be, external or internal service providers.

And then the last role, which is important is, we call it the management role, which are typically central digitalization teams or CDO type of organizations which supervise the whole thing. So which people are using the system? Which are granted access to it? Which type of solution elements are in the system so that you can build a solution with that? So it's overarching, I would say digitalization and strategy view. And to be honest, looking at these third at the end parties, it's more or less true for all the customers that somehow these organizations need to work together. End user and production or in manufacturing. IT side of teams or digitalization, coding, service teams and strategy or digital CDO type of organizations, in order to bring something to life. And that's what Axulus also manages as a standard from the engineering system that these parties work seamlessly together on concrete use cases and installations.

Niall Sullivan, Senseye: And do you often... You mentioned at the end there sort of C level, but more on the sort of digital IT side, obviously very similar, lot of similarities there between us and Senseye, but is there other input from other members of the C level or are they just more kept informed rather than being involved in the day to day running of these type of projects? I'd imagine so.

Florian Beil, Axulus: Yeah. I mean, look, I think... It's an interesting question. I think the C level is also involved in the sense that they are the final decision makers on the whole thing and what get installed. I think one of the key features which Axulus provides, and it would be the management or the CDO type of role I described before is, that as we have in Axulus also the value design of a solution, part of the stack, I would say, we also have a feature where you can, before you install a digital solution in the shop floor, you can simulate the impact on your target KPIs or the value add.

Which then at the end of the day would also provide them some sort of simulated value add you generated out of the system with all the different installations. And I think this is also where management, the C level, comes into play because at the end of the day, what they need to be interested in is, okay, digitalization is important, we got that. We enable the team to use digital technologies to become better, but what is now my quantified value add I generated? So did I save a million or 50 million? Or how much more OE do you get out of my solutions installed?

And I think this is one of the key features that is interesting for the C level to get involved in monitoring. So who uses which use cases? Which templates? Where are they installed? And then also getting out of the system. And I mean, an indication of what the value generated out of the solutions is at the end of the day across the organization so that they can also prove to the outside world, hey, this is what we've done and this is how we measure it. So I would say they are important, of course. I would rather keep them in the thirdth bucket. So management, and then there's a special focus on having a look at what is being done and especially, what is the return on that? And Axulus can provide a simulated value add out of that. So it's one of the features we have in the system.

Niall Sullivan, Senseye: Okay.

That was the third part of our series, looking at the challenges of implementing Industry 4.0 solutions. I hope you enjoyed it. From our conversation, Florent is clear that end users should be allowed to design and take control of their technology deployments to allow them to fully embrace the benefits of the industrial internet of things.

Please subscribe by your favorite podcast provider if you'd like to be notified about future episodes. And it would mean a lot if you could let us know your feedback by leaving us a review. You can find out more about how Senseye can reduce unplanned downtime and contribute towards improved sustainability within your manufacturing plants by visiting Senseye.AI.

Thanks a lot for listening.

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: