In this episode of the Industrial IoT Use Case Podcast, I talk to Frank Seiferth, the CEO of SEITEC and also seioTec GmbH. Frank studied electrical engineering/automation technology in Jena and founded SEITEC directly after his studies in 1998. Over 20 years, he has built the company from a “one-man show” to an established and innovative company in the field of industrial automation. He is a strong networker and maintains numerous contacts with universities and research institutions. For several years, Frank Seiferth has been driving the topic of digitization and IoT/IIoT specifically for medium-sized businesses. In order to focus even more on these topics, he and Kay Hölzemann founded seioTec at the end of 2018 as an IoT spin-off from SEITEC.
The direct view from medium-sized companies helps at this point to understand from practice which way the company SEITEC has gone and which challenges go along with it. In this context, we talk about the spin-off seioTec GmbH, which focuses on the topics of Industry 4.0, digitalization and IoT solutions for customers. In the past, SEITEC has constantly tried out the latest technologies and was already a pilot user for connecting a plant to Siemens MindSphere in 2017. The new spin-off benefits from more than 20 years of industrial, automation and electrical engineering know-how and connects the OT world with the IT and IoT world. Here, especially the consulting part and methodical workshop concepts have developed strongly.
Frank knows similar applications from the heavy machinery environment and the cement industry. Here, the company seioTec offers the customer an “all-around carefree” package, which ranges from the retrofit concept to the selection of gateways and app development. Frank explains that seioTec relies on established gateways here and is constantly on the lookout for stable partnerships. Depending on customer requirements, integration takes place in the desired cloud platform (PaaS). Cooperations exist here with major providers such as Microsoft Azure, Amazon Web Services (AWS), Google or Siemens MindSphere.
Industrial IoT app stores already offer the ability to scale developed software-as-a-service solutions for customers alongside their core business. The main difference is that connectivity to the field is required in addition to the classic software, which can be developed using the so-called DevOps (developer-operator) process. Due to the heterogeneous structure of old and new industrial machines, sensors and other data sources, consulting in the context of retrofit, security and connectivity concepts is relevant. Frank states at this point that with each use case developed, there will be more and more repetitive sub-functions of the applications. These can be standardized and scaled over time.
Via the process-driven optimizations through EDGE computing tied to Industrial IoT solutions, we come to apps for manufacturing midmarket. In the following course of the conversation, Frank explains a specific and scalable solution of seioTec in more detail. Customers from smaller medium-sized businesses can use the so-called “mini MES” app to obtain transparency on machine data, efficiency of selected key figures (OEE) and causes of stalled production. On an EDGE device, i.e. On Premises (OnPrem) through usage and licensing model for server-based software solutions, the data on missing raw materials or delayed machine starts is pre-processed by the employee and transferred to the cloud in a dedicated manner. Often, this use case is relevant for customers who do not have their own IT department to support IoT topics in detail. Output for these customers the “mini MES” solution from seioTec, which follows an EDGE-to-Cloud combined approach. This allows management to compare data from production at multiple sites and optimize based on cause.
Use Case 1 | Glass recycling manufacturer | Long-term monitoring of process data
There is also a solution for OEMs with complex supplier structures to save planning time. The use case comes regionally from the Thuringian Competence Center in Ilmenau and is implemented together with an SAP integrator. The customer is a medical device manufacturer (OEM) that works with multiple service providers and suppliers across multiple locations. The OEM’s problem is that the order and delivery date cannot be directly tracked. Often, reconciliation is done via email, but there is no overall picture of where which assemblies of the medical device are located. Transparency about the location, i.e. whether the part is in final inspection or still in assembly, would optimize planning times here. Frank explained that he initially believed that suppliers did not want this data transparency. However, as the project progressed, it became apparent that the OEM was developing stronger customer loyalty and trust through cross-manufacturing planning. The added value is obvious: Planning processes can now be carried out dynamically by the OEM and times can be optimized as a result. An integration of ERP and human resources of suppliers and sub-suppliers, as well as tracking of parts via RFID lead to transparency. This use case becomes particularly interesting when feedback from live data from the shop floor is included. Finally, a “digital shadow” is created as an image of the production depending on the suppliers.
Focusing on the technical solution at the start of an Industrial IoT project often does not lead to the goal. Frank explains what other mistakes seioTec GmbH has learned from in the last 2 years. The connectivity and detailed functions should be formulated only after the so-called user experience design. The operator can use a simple MVP to determine which functions are useful at all and which are not, right from the start of the customer journey. In this context, the MVP is the dashboard developed in the front end of the app, which is already operable, but does not access any real data. This is intended first and foremost to quickly and easily show the first minimally functional iteration of the software. After the feedback from the participants has been incorporated, the app can then be further developed. It should cover the minimum functional requirements and ensure relevant feedback for action from all relevant stakeholders.
The company seioTec is also involved in a research project, which is composed of a consortium of different companies. The project is called SPAICER and consists of a broad network of over 40 associated partners. The BMWi part-funded consortium is developing a data-driven ecosystem based on lifelong, collaborative and low-threshold Smart Resilience services by leveraging leading AI technologies and Industrie 4.0 standards. The goal is to foresee disturbances (anticipation) and to adapt production planning to active disturbances in an optimized way at any time (reaction). As an example, Frank cites tools used in production. Questions such as, “What happens if I don’t have my tools in stock?” or “Can I predict tool breakage and how can I better prepare next time?” are sifted through. In this consortium, SeioTec is responsible for software and sub-application development to provide resilience management for companies on multiple levels. The goal is to have a sub-ecosystem emerge where other companies, who in turn have developed AI methods for weather forecasting, shipping lanes, etc., can also bring them in.
At the end, Frank gives us his view on the future and the forecast on IoT technology in the next 5 years. The medium-sized mechanical engineering sector has become aware that digitization is a huge topic. Pure mechanics is becoming more and more comparable. Software-driven approaches and artificial intelligence are creating opportunities for advantages in the market. New digital services will emerge, and Frank sees great opportunities here for medium-sized companies as well. In the field of AI, it will become increasingly important to create reproducible approaches. Big Data is becoming Smart Data, and these approaches need to be communicated in an understandable, broadly simple and applicable way. AI must be comprehensible, interpretable, and reproducible by subject matter experts. This is also where boundaries of ethics and areas of application will emerge in the coming years.
Are you interested in a cooperation with seioTec GmbH?
Contact Frank directly via Linkedin!
More about this implementation partner
seioTec GmbH is the IoT spin-off of SEITEC GmbH. SEITEC GmbH has been supplying innovative solutions for manufacturing and process automation, as well as industry-related software developments, for over 20 years. The range of services extends from conception to complete project management and worldwide commissioning. For some years now, the interdisciplinary team has also been developing more advanced solutions in the context of digitalization and Industry 4.0. Since 2018, the company’s own IoT spin-off seioTec GmbH has combined 20 years of experience in industrial automation with the many years of IT and WEB technology expertise of co-founder and CTO Kay Hölzemann. seioTec develops customized IoT-based digital business models and services for its customers and accompanies them through all phases of the digital transformation, from consulting and ideation workshops to app development for established IoT platforms, as well as their operation and service.
Podcast Transcript Episode 13
Ing. Madeleine Mickeleit |
Hello Frank and welcome to the IoT Use Case Podcast. Yes, I am very happy to have you here today. How are you and the team doing at the moment?
Frank Seifert |
Yes, hello Madeleine. Thank you very much for the invitation. We are doing relatively well, so is the whole team. We are, I have to say, fortunately also quite spared from this whole issue of corona crisis. Yes, we can still continue to work quite well at the moment and have also returned to relatively normal office operations in the meantime. Some of our colleagues are still in their home offices, but we are able to distribute this a little bit among the offices because of the different locations we have, and in this respect we are able to work well and the team is doing well.
Ing. Madeleine Mickeleit
I’m glad to hear that. Very good. I’ve already explained a few points about you and SEITEC or seioTec in our podcast intro beforehand, but I think it would be quite good if you could personally say a few points about yourself and your current role at SEITEC.
Frank Seifert |
Yes, my name is Frank Seiferth. I am 48 years old. I studied classical electrical engineering automation technology and founded SEITEC in 1998, directly after my studies. I was already working on the side and when I had my diploma in hand, I started my own business and then founded SEITEC accordingly. I moved around the world there for a few years doing PLC programming classically as a freelancer in PLC and SCADA system programming, literally. I have performed many worldwide start-ups. Meanwhile we are at SEITEC in this automation area, 18 engineers and computer scientists at 2 essential locations, we here in Thuringia, I am sitting right now here in the beautiful Königssee, not at Königssee, but in the middle of Thuringia and here I founded, here is also our headquarters. We still have an office here in Erfurt in Thuringia and a branch in Bietigheim-Bissingen in the Stuttgart area. And meanwhile with this team we offer complete solutions in the field of industrial automation and additionally industry-related IT and software solutions. We also see ourselves essentially as a partner in mechanical and plant engineering. We have some regular customers from this field and solve their tasks in the field of automation. But we have also always had a focus on IT, which is why we will come to that in a moment, we have always gone beyond classic automation and have always had a big focus on research and development, new technologies. Yes, and so we have now also arrived in the age of digitalization and automation.
Ing. Madeleine Mickeleit
Okay. Exciting. For me, it would also be exciting at the beginning, as you already mentioned at the beginning, you take on automation topics, you are a partner for machine and plant construction with an IT focus. For me it would be exciting to show this difference between SEITEC and seioTec GmbH. What are the classic tasks that you take on in detail at SEITEC or seioTec GmbH? And how advanced is perhaps also your vision in the direction of IoT or Industrial IoT in that context?
Frank Seifert |
Yes, that’s quite exciting because we have actually gone through our own digital transformation as well. The story is not quite over yet. So 20 years after SEITEC was founded, I founded our own IoT spin-off together with Kay Hölzemann (CTO seioTec), the current CTO of seioTec GmbH, in 2018, seioTec GmbH with the IOT nicely in the middle of the name and that’s where we are now focusing all these topics of digitalization and IoT. That’s just come that we’ve always tried to use the latest technologies and we’ve already gained experience with IoT applications since 2017 approximately. In 2017, we had already connected a plant in the USA, for example, to Siemens MindSphere as one of the first plants, I believe, to be an external partner, and we operated it on the side for a while and noticed that this is really a topic for the future. And we are now focusing on these IoT topics in our own IoT spin-off seioTec GmbH, where we also benefit from the fact that we have more than 20 years of knowledge of the shop floor, of various machine processes, and really combine these worlds of OT and IT in a super way, I think.
Ing. Madeleine Mickeleit
Yes. I was just about to say that these two worlds are merging there with you from the automation world into what is now the IoT world. What are some of the classic tasks that customers are asking you to do now with seioTec? How do you positioned there?
Frank Seifert |
So, that’s a very broad field, I must say. That was also one of the reasons why we founded seioTec GmbH, so that we can now focus on more than just these purely classic industrial topics. A typical case, also one of the first cases and still frequently with the most use cases in this area, is mechanical engineering, for example a German company from the mechanical engineering environment, but 70-80% of its machines are located at its customers’ sites worldwide, and topics such as transparency of the data when the machine is actually located at the customer’s site, proactive service concepts, and of course topics such as “what does maintenance mean for us”, which we also address somewhere in the end, come up. So there are always many topics from mechanical engineering, which are also somewhat similar. Every mechanical engineer has his own focus there, of course, but there are many points of overlap. At the moment, these are most of the use cases in this area.
Ing. Madeleine Mickeleit
Okay. Now you said that you yourselves have gone through the digital transformation internally, so to speak. What were your challenges in dealing with this issue, shall I say, or tackling it at all? So how did you do it?
Frank Seifert |
So, it already starts that you have the whole team and that you have other employees. At SEITEC, we tend to be the classic electrical engineers and automation specialists, and now we are using web technologies. Even if we have computer scientists at SEITEC, they are more concerned with Windows software and database topics, and now the focus is much more on mobile applications. The topic of UX design (user experience) is also moving much more into focus. This means that there is already a big difference in the team and also in the processing of the projects. This is already a change in our company or a change in thinking from the direction of the management, also in my own case, because I am more of a classic engineer, that you have to approach projects in a completely different way right from the start of the first customer meeting.
Ing. Madeleine Mickeleit
Mmh. How do you do it in the classical process? Because at the end of the day, you also have a certain logic that has become established in processes for tackling projects. Now in the IoT environment, it’s more the agile, you have new ecosystems somewhere that you need, partners that you access. How do you go about it and how do you manage to bring such agility into your everyday work?
Frank Seifert |
This actually starts much earlier. The development is just at the moment in the current phase yes the second step only. Currently, we are much more active in the areas of consulting and first-step advice, because customers or interested parties in this area often do not know exactly what they really want. It varies greatly, but many have the topic of digitization and IoT somewhere on their agenda. The technologies are all there. We have a whole new technological age here. All the technologies are there now, and you have to think about what to do with this brave new world. And that’s why it starts much earlier now with very intensive consulting processes, workshops, ideation workshops, co-creation. And this is actually also an issue that we never really had in classic automation. Consulting yes, but not in this intensity. That is the first difference. And in software development, if you can get into it, similarly. We then really go a) you have a lot of things like a proof-of-concept, going over an MVP or going “prototyping” and then very soon our customers are trying, and that’s also a success strategy then from the customers at least, getting experiences back from their users. And in the classic automation world, you often have a requirements specification phase. Then it’s clear, you plan for weeks and months, you create software, and then you go to a plant and put it into operation. And there are now also completely different development processes.
Ing. Madeleine Mickeleit
What approaches and ideas do you see from your customers? You just said, often the customers themselves don’t know, let’s say, 100% what they want, what they need. When you go into a consulting project like this, what use cases do you see at your customers that work in the IoT environment or at least, I’ll say, where you already have ideas that have been approached in the context of proof-of-concepts in medium-sized companies?
Frank Seifert |
There are many use cases that we are talking about, but what is currently similar from the perspective of mechanical engineering, which I mentioned at the beginning, and which is proving to be successful for many, is first of all this topic of obtaining transparency about my machine data and a topic that relates to maintenance activities or the topics, These are relatively successful topics for customers because, once I have more data, I can build up further topics based on this. And often we have noticed at the moment, the theme of maintenance is the first entry.
Ing. Madeleine Mickeleit
Mmh. Can you do that, maybe just give an example? I mean predictive maintenance. Do you have a customer who has a specific problem that you are working with, or what does predictive maintenance mean for your customer in this specific case?
Frank Seifert |
So with predictive maintenance, I’m still a little bit cautious from the word formation, because we have some experience and knowledge in the application with AI methods and models. With these products, we are not yet really in predictive maintenance, I have to say that, but that is where the journey will take us. At the moment with a concrete example: we have at a customer, these are really heavy machines in the field of mining, originally a bit cement industry, mining or glass recycling, so such larger grinding plants, relatively heavy machines and also wear-bringing processes and this is the typical case, a German company and even 80% of the projects and plants are distributed somewhere worldwide. And the customer has usually had a one-time deal, the machine was sold. They might have had a remote service contract and also didn’t know exactly how does the machine work and that’s where we really went through a workshop process via ideation workshops and first worked out the challenges together. What are the challenges of the customer, i.e. the machine builder at that moment, and what goals would he like to set for himself? We sometimes say that it can be a “make a wish” process, and of course you have to work out what is feasible and what fits technologically, but also within the budget. And I could give you a couple of examples of what issues were addressed there. And that’s a typical process as a starter, which is really 2-3 workshops.
Ing. Madeleine Mickeleit
Okay. Talking about glass recycling now. How is such a system structured and what kind of data, let’s say, data that I need or would need in such a field in order to perhaps also align my service with it? Yes, maybe also predictively monitor this process as well and maybe give the customer hints on operation, etc. Or what are, as they say, the “pains” of the customer?
Frank Seifert |
Yes. So you have to imagine that in these plants, broken glass, which typically ends up in these glass containers in Germany, is ground into small pieces, almost into glass dust. It’s actually like cement at the end, so to very fine dust that is ground up. And you can perhaps imagine that this process, the abrasive materials of which is very wear. So you have silo technology and these conveyor belts that convey broken glass into the plant and then it is ground in a circuit with classifier technology for such a long time that the fine material already goes out and what is not fine enough is fed back into the circuit until I really have this fine glass dust at the end. And I have hydraulic units there with oil monitoring, with filters in them, I have various filter systems for the exhaust air, and of course I have these rollers between which the glass is ground, which are pressed together with high hydraulic pressures, and I have large loads on gears and on motors. These are actually many things that we monitor there, i.e. vibrations on the engines, monitoring of oil quality, of oil from filters and these are actually just typical measured variables that are also repeatedly discussed in many of these plants, which we also monitor there. Of course, we had a control system with a local HMI system, of course, where I am also notified and have a certain logic in there already, but one hopes here of course now through the long-term recording of data and also recognize correlations, how does my raw material, that is an essential point, I could not record so precisely before, how does it affect the composition of my raw material or also, for example, ambient conditions in the sense of humidity in the ambient air plays a big role here, because otherwise it comes to adhesions. And if I now collect and evaluate this data via really more long-term data evaluation, I will hopefully also be able to handle real predictive maintenance. At the moment, these concepts are that we really first also carry out limit value formations and inform in good time before the local HMI responds, but also already have certain intelligences to carry out projections of running times of switching mirrors with a weighting in order to first implement this preliminary stage, let’s say proactive service aspects, because for real predictive maintenance we also need a bit of data analytics. That’s a phase we’re in right now. I can’t report any results there at the moment, that will be the next stage.
Ing. Madeleine Mickeleit
Yes. Okay. Super exciting. How do you bring your automation knowledge to the table? What does such a, shall I say, connectivity solution look like then to some extent? So I’ll probably have the mechanical engineer who has an EDGE device like that running AI algorithms like that. What does such a concept of yours look like then?
Frank Seifert |
On the one hand, it is an advantage for the machine builder that we can talk to them at eye level, because we have already experienced many different processes in our 20-year history of automation and don’t just talk with these IT glasses or don’t just have the IoT glasses on. This is often very important in order to understand the processes. In the direction of connectivity, of course, we always try to find the optimal solution. An example would be: I already have a PLC at the plant in this example, of course, and at the moment we still have a small industrial PC installed at the plant, do a certain amount of data preprocessing, data collection there and then send consolidated data via a connectivity, in this case it was Siemens MindSphere in the first example, and then send this consolidated data to MindSphere.
Ing. Madeleine Mickeleit
Yes. When you mention Siemens MindSphere, for example, which partners do you work with in such a case and who takes on which part? This also includes an app, connectivity concept, etc. Which area of the IoT layer model are you taking over, or up to what point?
Frank Seifert |
Without wanting to exaggerate, we actually offer our customers a complete package of solutions. With partners, of course, but we see ourselves as a solution provider. We try to offer the customer a holistic solution with this holistic approach, which we have precisely in this OT-IT connection. So if we talk more about seioTec more about these IoT topics, for us actually the biggest part is to develop the app in the end. That’s the biggest value-added process for ourselves, developing the app and even running the app for some customers, providing them with an all-inclusive package. In the topic of connectivity, we naturally try to rely on established gateway solutions. We don’t want to reinvent the wheel here, so we work with partners there, mostly in the sense of component suppliers, we are also constantly looking for new partners there for EDGE devices, for gateways, for example. Of course, we rely on the established platform providers. We had already mentioned MindSphere, we see ourselves as a neutral partner, we also implement applications natively on Microsoft Azure or on Amazon Web Services. That is, we actually try to work out the best customer solution.
Ing. Madeleine Mickeleit
Just quick interjection. You see, because you have now addressed the applicative part, do you see a scalable approach for you in the sense of, you might have an application for a glass recycling topic. Can I also do that or do you want to sell this app as well, as you know it from the private environment or is that customized or what approaches do you see there in that area?
Frank Seifert |
Yes, definitely. After all, we have 3 business models in the IoT environment. One starts in consulting and ends in a customer-specific application. This is a typical project business actually. The second is that we would of course also like to develop generic apps, and the first approach also exists, namely precisely from this maintenance environment. We are currently casting these core functionalities from our maintenance application experience into a generic maintenance app, which we hope to be able to sell in the future on established app stores in this industrial context. Where the customer just says, I’m going to download the seioTec Maintenance App and I’m just going to use it, and of course that’s also our focus, and I’m convinced that the more applications are implemented in customer projects, the more ideas will come in the future. Maybe always for small sub-functions in the first step to develop generic apps and generic solutions. That will also be a clear focus in the future, because quite honestly, I think there are still too few use cases overall, which makes it difficult to work out these functions that everyone can use, but that is clearly our focus.
Ing. Madeleine Mickeleit
Yes, if I now look at the area of medium-sized businesses with these apps, does that also go in the area of a mini MES app or something similar? Because we often know that even medium-sized customers, some may have an MES system, but the multitude may not, can I then also solve something like this via an IoT app? Is that the way it goes or is that something else again?
Frank Seifert |
Yes, good keyword. Yes, is a second area in addition to this classic mechanical engineering, which we have noticed in the many sales discussions, also customer discussions of the last year, that a second field, just as you said, the middle class or smaller middle class, because we have just customers so manufacturing companies, which have several CNC machines and maybe a few other processing machines, but is not a larger production. They have an EAP system when things are going well, but rarely an MES system, but would like to have visibility into their machine data in their local production environment. Of course, these are other reasons. There, it is often about OEE topics, i.e. efficiency ratios, also there locally, of course, predictive maintenance topics, can I already predict things, why is my production flow stalling, is it due to a lack of raw materials, is it because my setter did not restart the machine in time or or. So there are many things and the advantage is with these technologies, we’re really talking about EDGE technology in combination with IoT, that this smaller customer, who doesn’t have an IT department anyway, doesn’t have to worry about all these IT issues because these are just technologies now, it’s running somewhere on an EDGE device, browser-based, he doesn’t have to do a big IT installation and worry about Windows Update. For this reason, this technology is very interesting for the. And that’s just a second customer area as well that we’re continuing to focus on.
Ing. Madeleine Mickeleit
Yes. For many smaller customers, this is insanely exciting. Now I just wanted to, you just mentioned that, but then that’s a solution that you run on premises, so on an EDGE device. This is not a classic IoT case where you say, “I’m going to build a mini-MES in the cloud. Or are you already using cloud resources in that case?
Frank Seifert |
Mixed. The first approach is On Premises and since you mentioned earlier also the topic of ecosystem and that is actually a good keyword at this point, here we are also tacting for example companies that are addressing the management of EDGE devices, central software roll-out and similar topics. At the moment, there is an incipient cooperation from your IoT podcast with the company CTO Cybus GmbH, with whom we now want to start a first project, because they have this connectivity on the store floor.
Ing. Madeleine Mickeleit
Oh, nice. Yes, that’s right, Peter had already spoken about a solution for MES middleware in the podcast.
Frank Seifert |
Exactly. So, of course, we are also trying to find good partners there in different ecosystem areas to find the best solutions. You can’t do it all yourself in this environment. Clearly. We then present ourselves to the customer as a total solution provider, but we also try to solve this through good cooperation and partnerships. And yes, a lot of it is on the EDGE device at the moment, we just have a case where a customer has, for example, 2 production sites, and there is just the great mixture, yes each site locally is evaluated on the EDGE device, but we have just through our know-how also the possibility of consolidated data of both sites in the cloud, so on the IoT platform for him. So he has a transparency from the management also of certain data about both sites.
Ing. Madeleine Mickeleit
We had now focused so strongly on the machinery and plant engineering sector. Do you also have use cases or projects that take place in a completely different industry that you might not have had on your radar before? I don’t know, are you active there?
Frank Seifert |
Yes. So it’s really exciting to see what use cases are being talked about in some cases, and a very interesting one that I never had on my radar has now come up here in Thuringia, also through a cooperation with the Industry 4.0 in Ilmenau and an ERP system integrator that use IoT platform technology to implement what we used to call cross-production planning, i.e. an OEM with several suppliers can synchronize its ERP planning with the sub-orders that it gives to its sub-suppliers and thus obtains an overall picture of its total load including the sub-components, so these are entire assemblies that he has delivered to him, and that’s where we connect entire production sites via the platform. And so that’s a really exciting topic that I wouldn’t have thought of before, but it came out of conversations like that and it’s really quite an exciting topic.
Ing. Madeleine Mickeleit
Yes. Then what kind of customer is that when you say OEM? Can you name that one? Or is this a secret, i.e. “restricted project”?
Frank Seifert |
Yes. At the moment, this is still a bit of a research project. So you can say that quietly, this is a medical equipment manufacturer who has several suppliers here in Thuringia and the problem just had so far, so something runs nowadays still these subcontracts about, yes an order is placed and delivery date is specified and then further arrangements are made by phone or e-mail ………. in time the part is there. And now the following transparency is practically created, that the OEM as the order giver can really see, is the part really already in final inspection at my subcontractor or still in pre-assembly. And I’ve always thought that no one wants that, because of course the supplier may also be afraid of showing this transparency, and now it turns out that both say we have added value, because the trust is greater and of course I have to show a certain transparency that did not exist before. But for the OEM, it is also an issue to continue to work with the Thuringian company, perhaps, because the basis of trust among each other is significantly strengthened by this, and what is still very exciting in the project is that we, at the beginning, only talked about this ERP data and with the technologies that we now have in the IoT context and also with the automation and shop floor know-how that we have, and also to integrate connectivity solutions somewhere, we can even find dial-up time data of the machine in this project, integrate it with this data pool of the actual planning. So you move away from a stringent pure planning track, you even get the feedback from the shop floor. And then the topic becomes even more exciting.
Ing. Madeleine Mickeleit
Yes, i.e. you can, what you just said, where is my part, is it still in the final assembly or somewhere in the final inspection, i.e. you connected the data e.g. from an assembly station, where you then go via piece counters or how does that work then or which variable parameters do you need there from the field?
Frank Seifert |
The first step is to take another step back, so to speak. There is ERP planning, i.e. OEM has somewhere planned a sequence for an assembly of a total analyzer and a part of it is with a subcontractor and he has his own ERP planning, when is the part planned in my production. And now we can, that is even very few data only, we know where this assembly is located via RFID, simple RFID tracking in the production and can therefore first only play back and that is already a great added value, is the part now still in my pre-assembly or has the part arrived in the final inspection. After all, that’s enough to make the planning process of the higher-level OEM company much more adaptable again. That’s what this is all about. We have a very innovative ERP integrator on board in this consortium, which is the university in Ilmenau and an ERP system integrator and we, as I said, I can’t say too much about it because it is still research, it is already prototypical, it has already been implemented here in Thuringia, but I can’t name any names at the moment.
Ing. Madeleine Mickeleit
Yes. Of course. We may be able to submit it in half a year’s time. Depending.
Frank Seifert |
Right, what you’re saying right now. We are also checking which data from production can be usefully incorporated, because the know-how of the ERP integrator is to recalculate these planning processes completely dynamically and automatically. So recalculating production planning sequences on the EDGE device. We are talking about a digital shadow, not a digital twin, which would otherwise be a blurred image as a digital shadow, i.e. a digital image of the complete production planning and that is adapted depending on many parameters, personnel resources, among other things, which is not so much our topic, but for us the interesting thing depending on the suppliers. Where is the part currently at the supplier. And that’s really one of the most exciting projects, if I look into the future, that we’re working on right now.
Ing. Madeleine Mickeleit
Now you have seen several projects, IoT use cases, etc., also in the midmarket. Do you have any practical tips where you say, hey, we should pay attention to this in advance, or we made a mistake, we could have saved a little time, or something similar? Do you have an example from the project without naming anyone or providing the know-how at this point?
Frank Seifert |
Yes, I would put that as a general example. From the project experience now of almost 2 years in the implementation of such projects. We focused too much at the beginning also on the complete technical solution also at the end yes early project how is the connectivity or what do we have for functions and what has turned out also from the point of view of the later user or our customer, we now always start very very early first only with UX design. We start with UX design.
Ing. Madeleine Mickeleit
So user experience in the sense of how it looks in the end for the operator, so to speak.
Frank Seifert |
Exactly. Because with this, the complete, I say, costumer journey is actually mapped, even if the function behind it has not yet been developed, and the customer can get on board with his sales department or his other internal stakeholders at an early stage, or can even get on board with sales, which is already a functional mapping, so you can really operate the browser, for example, no results come back, but I can navigate. And that’s very good at the moment, that we’ve already pushed it quite far, sometimes to our disadvantage, because the customers then already think, oh you’re already finished, everything’s already developed, why does it now cost X amount, because it already looks so close to reality, but it offers this mechanical engineering sector, which for example also has a very hard time with such new topics, the opportunity to approach its customers early on and also, as I said, to bring stakeholders on board. And then we develop functionalities in parallel with short sprints. So that is a really very important experience that we also had to make ourselves, that it works much better this way.
Ing. Madeleine Mickeleit
Yes. Now you are also involved in other projects. Maybe we can address those very briefly. I think you also have a research project like that, which you do on the side, I’ve seen. What’s the deal with that and how else are you active in other industries as well?
Frank Seifert |
Yes, we are also involved in a very exciting research project in this context IoT. This project is called “Spaicer”, you can also find it at spaicer.de, I think, I’m not sure, but you will add the link. There was the AI initiative of the German government at the beginning of last year, so the AI innovation competition, where several consortia applied to further develop AI technologies and we are fortunate that we were selected in a consortium led by the German Research Center for Artificial Intelligence, DFKI, here to the final round and then even at the end of 2019 from an initial, I think, 85 applications or even 100 applications and there are now, I think, 13 projects funded and here it’s about resilience management for industry. That is, how can companies be strengthened and made resilient to external influences? There are several topics being looked at there, so maybe the easiest one to explain. If I look at shop floor resilience, we’re talking about micro-resilience here in this project, for example, a typical approach for predictive maintenance would be what happens if I don’t have my parts in stock. Or can I predict tool breakage? How can I better prepare for that next time? That would be micro-resilience like that, but that goes all the way up to issues like last year low water levels on our rivers and lead chains were interrupted, BASF couldn’t have raw materials anymore and if you had been able to do that through methods of AI and, of course, because of that, on platforms including weather forecasts, etc., you could have predicted that in a way and you could have done production planning or supply chain planning differently as a result of that. That is described in a few words, so if you are interested, take a look at the homepage of the project, so roughly what it is about and we are responsible in this core consortium for this entire topic of software development on the platform. In the end, applications, probably also several sub-modules or sub-applications, are to be developed in order to be able to offer this resilience management as software applications to companies at these different levels. That’s why there will certainly be several applications that will be created there in the end, and the goal is really to create a sub-ecosystem where others can also import AI methods for weather forecasts or for shipping routes or similar, so the goal is to create a sub-ecosystem in this context. And, yes, the project is now running for 3 years with a very great consortium, where very well-known industry players are working together with us. And of course we are pleased to be part of such a project, also the latest findings from research and development, university, landscape, research institutions, yes, for us too, of course, this is somewhere an incredible increase in knowledge and know-how.
Ing. Madeleine Mickeleit
Yes. Like you said, I’ll link all these use cases that we’ve discussed now accordingly and also the “spaicer project” that you can also find the website if there’s interest, otherwise there’s the connect to you anyway. Now looking into the future, even before I finish now, what do you think will become more and more important in the next 5 years, what will we see in the IoT market, maybe also in the midmarket especially? What do you think will come in the next 5 years?
Frank Seifert |
I believe that right now, of course, we are talking about digitization in a crisis-ridden context, but I have already had the good fortune to have 1 – 2 customer meetings again and I believe that the medium-sized mechanical engineering company has really become aware that digitization is an issue that it must take up. And pure mechanics, such things are becoming more and more comparable in mechanical engineering, for example. One can still stand out through software applications probably more and more with a share of AI here, can offer new software-driven topics to its customers, where one can have advantages on the market. I also believe that there is really, you could tell, a lot of people were interested in the last 2 years, but there was a lack of more concrete ideas and I believe that something will happen in the next 5 years, also due to the growth of ecosystems, you had a lot of very interesting companies in your podcast, from connectivity from the smallest sensors to prices that you can then also present accordingly on the market and with this, many digital services will emerge that we are not even aware of yet and I see great opportunities for us. I really see a future there. In the field of AI, it will be important to have explainable AI, also a reflection that I can give, that many confirm. AI in terms of Big Data to Smart Data and I start something with your data and it ends up coming out with a great result, if I can interpret that correctly, it’s not going to meet with acceptance. After all, we think about the topic of AI in many projects. However, it must still be comprehensible in some way by the expert. He can’t understand it down to the last detail, but it has to be reproducible, what did the AI really do now. And I think there will be a lot of discussion in the field of AI in the next few years, how do I use AI, where are the limits, also ethical limits, and yes, I think that will keep us all busy in this field for the next 5 years.
Ing. Madeleine Mickeleit
Yes. Thanks so much for your insight here. I also find it insanely exciting viewed through your lenses. I mean, you come from the automation world and are of course on a par with many, let’s say, listeners who now also come from the most diverse industries, who are facing exactly these topics and it is really super exciting what you have created there, also with seioTec. We have now had various use cases from heavy machinery mining, to the topic of the medical device manufacturer from EAP integration of data from suppliers, to ERP, how do you say, mini MNS systems. I will link that accordingly as well. And if there is interest, what is the best way to reach you? Is that then through seioTec via the website or LinkedIn or what is the best way to reach you?
Frank Seifert |
Yes, you can reach me via the seioTec homepage, and I am also active on LinkedIn. I guess, you’ll also link the contact information. So via SEITEC or seiOTec Hompage or LinkedIn are the best methods, all contact details are stored there. I can be reached there.
Ing. Madeleine Mickeleit
Okay. Super. Yes, Frank, then thank you very much for your time and for all your exciting insights here. I think that from your point of view, the whole thing has become a bit more understandable and you can also make the whole thing tangible in practice with the help of the use cases, which I’ll link to accordingly and which we’ll also publish. Yes, thank you again. I am very happy that you were here. Thanks for your time and enjoy the rest of the week.
Frank Seifert |
Yes, thank you for the invitation as well. It was very exciting and I also look forward to your future guests. It’s really a great IoT podcast. Great that you had the courage to start this IoT podcast as well. Yes, thank you.
Ing. Madeleine Mickeleit
Yes, very much so. Thank you. See you then. Ciao.
