DigitalituM Podcast Episode 9 - Sina Volkmann - FindIQ - Knowhow transfer in Field Service & Maintenance with AI

 Enhancing Field Service with AI-Driven Knowledge Transfer - An Interview with Sina Volkmann of FindIQ

Episode Description:

In this episode of the DigitalituM Podcast - At the Intersection of Manufacturing and Digital Transformation, host Markus Rimmele talks with Sina Volkmann, Co-Founder and CEO of FindIQ, a German startup revolutionizing the way knowledge is shared in field service and maintenance. FindIQ’s innovative solution provides "a service expert in the pocket," ensuring field workers have instant access to critical machine knowledge and troubleshooting guidance, even when senior experts aren’t available.

Key Discussion Points:

1. Origins of FindIQ:
   • Sina shares her journey from working in the machinery industry to co-founding FindIQ.
   • The impact of the COVID-19 pandemic on service delivery inspired FindIQ’s vision to bridge knowledge gaps when experts cannot be on-site.
2. Problem Statement:
   • Manufacturing downtime spiked due to limited technician availability.
   • Traditional knowledge management methods, such as paper manuals, lack efficiency in modern production environments.
3. Solution Overview:
   • FindIQ’s application is a mobile troubleshooting assistant and is accessible on smartphones, tablets, and PCs.
   • Built to be hardware-independent, the app guides users through troubleshooting by asking targeted questions based on symptoms and observations.
4. Key Features of FindIQ:
   • Service Expert in the Pocket: Real-time troubleshooting assistance replicating experienced technicians' know-how.
   • AI-Driven Knowledge Transfer: Structured, probabilistic approach to diagnosing and solving complex machinery issues.
   • Self-Learning System: Knowledge grows as technicians add new symptoms and root causes, ensuring the solution stays current and effective.
5. Customer Success and Real-World Application:
   • FindIQ has a diverse customer base, including machine builders and global production companies like Siemens.
   • The system improves efficiency by allowing operators to handle common issues independently while expert technicians focus on more complex challenges.
6. Future Outlook:
   • Sina discusses how technology is reshaping service and maintenance.
   • Emphasis on aligning people, processes, and technology to leverage digitalization in manufacturing fully.
7. Getting Started with FindIQ:
   • FindIQ offers a trial phase for companies to experience the solution's capabilities.
   • Interested North American companies can contact DigitalituM, FindIQ’s partner, for demos and support.

Contact Information:

• FindIQ Website: findIQ.de
• DigitalituM for North America: DigitalituM.com
• LinkedIn Sina Volkmann
• LinkedIn Markus Rimmele

This episode provides actionable insights into using AI to streamline field service and reduce downtime, making it essential listening for anyone in manufacturing and digital transformation. Tune in to understand how FindIQ’s AI-driven knowledge management system could help

Stay tuned for more inspiring conversations about manufacturing and digital transformation. Also, remember to follow and subscribe to the DigitalituM Podcast for exclusive insights from industry leaders and innovators.

We appreciate your likes and comments. If you feel you can add value to this podcast series and want to be our guest, send an email to Sales@DigitalituM.com

Transcript

Markus Rimmele - DigitalituM (00:01)
Hello everyone. This is the Digitalitom podcast at the intersection of manufacturing and digital transformation. Today's guest is Sina Forkmann with Find IQ, a new innovative startup from Germany. And they focus on the know-how transfer of field workers in the field with their magic app. Welcome Sina.

Sina Volkmann (00:31)
Nice to meet you, Marcus.

Markus Rimmele - DigitalituM (00:34)
So let's jump right into it. Tell us a little bit about yourself and how did you come up with the idea to get into a startup and found FindIQ?

Sina Volkmann (00:48)
Yeah, so basically you pronounced my name absolutely correctly. So I'm Sina Foghman, just turned 30 this year, which makes me quite somehow old for a typical German startup. However, I've been working in machinery industry for the last 10 years, I would say. So have always had ideas on how to digitize productions and processes.

Basically, the idea behind FindLQ was nothing like I woke up and had this magic moment of, I need to found a company. It was rather a real, I would say, rational problem we had in Germany originally, just due to the fact that during the COVID crisis we had from what was a 2020, I guess, right? So when it came up, there were a lot of service technicians of German machine manufacturers.

they couldn't travel. Right? So they were all stuck somewhere in Germany and all the machines around the world, they, I don't know, there was breakdowns and there was problems with the production, with the product's qualities and so on. And just due to the fact that they couldn't travel, the knowledge wasn't close to the machine.

And I guess the machine downtime cost increased by 70 % from 2020 to 2022 only due to the fact that the knowledge couldn't be somehow transferred by the people to the machines all over the world. And I was like, okay, I'm not really relevant to cope COVID. I'm not, I don't know, in medicine or can contribute anything to this.

global issue we had, but maybe I could help the industry to fix this problem. And then we had some, it was a research project going on. So they asked for young people in our area, so someone in the nowhere of Germany. And they asked, okay, is there anyone who has an idea on how to get this knowledge from Germany to the US, for example, because I guess that the borders of the US were closed as well. And I was like, maybe I have an idea, but I don't have the technical competence.

Markus Rimmele - DigitalituM (02:57)
Mm-hmm.

Sina Volkmann (03:02)
I've just the need, I understand what's the need of the customers. And at the same time, there was a colleague of mine, so he's my co-founder today. And he also applied to the challenge and he's a data scientist. So he's very, I would say into AI and has always worked with AI. And he was like, okay, I may have an idea on how to get this knowledge out of these people's heads and make it available just next to the machine.

And then it became projects. So you were funded by the government. So got some governmental grants. And after this research project, we had the first customers. And then we noticed that this problem could be big enough to fund a company basically. And two years later, here we are.

Markus Rimmele - DigitalituM (03:48)
that sounds all very innovative and actually fixes quite a problem. As you may know, I have, back in the days, I was a field service technician, field service engineer myself and struggled with the issues out in the field that the major challenge is get the information you need at the location you are as.

At least previously everything was still in paper, tons of paper binders, big German lights, altanels with manuals and drawings and all of that. And typically they're also not close to the machine in a factory. They're somewhere in a maintenance shop, which in big large factories is like a mile away from the actual equipment you're working on. And if you then walk into that maintenance room,

and you see this wall of paper binders, you have to find the right one and then look into that one to get that information. And if you're super lucky, your coworker just ripped that page out what you're looking for. And here you are.

Sina Volkmann (05:01)
Yeah, exactly. That's exactly the situation we had. And that works for small and medium sized companies when you can talk to each other. just imagine you have like globally producing firms. So you can't just go with your lights ordnance to the other person and say, okay, here's the information you're looking for. It's always like, it costs a lot of time to go there. It's not really efficient. And some of these guys then bring the wrong spare part or have the wrong

Markus Rimmele - DigitalituM (05:13)
Mm-hmm.

Sina Volkmann (05:30)
qualification when they go there, then they go back and mostly by plane or so. And I guess that that's, it's not only a pandemic issue. when borders are closed, but it's also something that's related to, I would say, inefficiency of how to cope with and how to treat resources, whether it's human resources or if it comes to the ecological part. And what we're currently noticing is that it's just exactly the same when you have a look at the US. So

demographics, people retiring.

Markus Rimmele - DigitalituM (06:01)
Yeah, yeah, I think overall we have the issue that machinery is getting more and more complex on one side. So it's kind of going up over time. Of course, like back in the days, a machine was operated by relays then with a basic PLC. And nowadays it's controlled by a computer and the software is somehow boxed up and you don't even access to it.

And then on the other end, the quantity and quality of field workers, either on the field service or on the maintenance side of things is going down as the paper boomers are retiring. Not only in the U.S., that's kind of like a worldwide phenomenon, at least in Europe and in North America. And for the young people,

manufacturing industry isn't that enthusiastic. They go in other fields, so we cannot fill the ranks. And that translates in a big gap, which to me translates in unblended downtime as a result for that. And therefore the whole know-how transfer is an important thing. And you guys have a solution for that.

So let's talk a little deeper in how this solution actually look like. So to my understanding, it's an app which runs on your phone, your tablet, your computer. So it's hardware independent, right?

Sina Volkmann (07:46)
Yeah, exactly. So it's just, it should be a mobile device at some point, because if you have a trouble shooting issue, something you need to identify and fix at the machine, sometimes it's just the PC next to the machine. then it's fixed, but should be something mobile. So smartphone, tablet, something like that. And then they just have like what we call it, the service expert in the pocket. Yeah.

Markus Rimmele - DigitalituM (08:12)
which is a good naming. you have kind of like the senior coworker who knows all and everything always with you at every location and at every time in the middle of the night and even on the weekends where you normally are not able to get anyone.

Sina Volkmann (08:37)
Exactly. So in Germany, it's like if you call it the service hotline for some companies, it's like, I don't know, night, a.m. in the morning, five exactly. Exactly.

Markus Rimmele - DigitalituM (08:44)
Monday to Friday, 8 to 5. And Fridays they close at 12.

Sina Volkmann (08:49)
Yeah, exactly. That's how it is. And if you just have a look at how high the export quota is of European or German based machine manufacturers, and in case they deliver to the US, I have no idea how they want to provide sufficient service capacity to US customers if they don't have the people available.

Markus Rimmele - DigitalituM (09:08)
Yeah.

Exactly. And getting somebody on a Friday in Germany, it's always a challenge as we are six hours behind. It's then Friday afternoon in Germany and everybody is gone in the weekend. And if there's a holiday, like on a Thursday or Wednesday and everybody takes the Brückentag, yeah, it's always a challenge.

Sina Volkmann (09:35)
Yeah, that's a German. Also, I've had a look at so different statistics. And what's interesting is the demographic structure in Germany and the US is similar, but the US is more productive than we are. So basically, we do the same, kind of effort, but it's not resulting in results in the same output, would say, or outcome at the end. I guess that...

Markus Rimmele - DigitalituM (09:45)
Mm-hmm. Mm-hmm.

Mm-hmm.

Sina Volkmann (10:02)
If you have a look at also trends going on here and people who don't want to travel, but also want to stick to usual working hours, it's just not possible if you have like a three shift production, if you have production somewhere else, I don't know, on the globe, then how to provide sufficient service 24-7 in I would say 20 different languages if productivity goes down in Germany. So this is an additional issue. What makes it like...

Markus Rimmele - DigitalituM (10:19)
Mm-hmm.

Sina Volkmann (10:31)
yeah, even more important to somehow, I would say, multiply this knowledge and try to enable as many people as possible to somehow get the production up and running, right?

Markus Rimmele - DigitalituM (10:43)
Yeah, I agree. Yeah, and with this whole know how transfer topic, I know other systems on the market, which are based to take existing data and data structures. And then with the magic of AI, get the input out of it.

The issue what I see in a lot of machine building companies on the small medium side of things, which in Germany there are thousands of them, is that the data structure is not really there as the data is in the head of the experienced people. So what do you guys do different and how do you get that tribal knowledge?

Sina Volkmann (11:21)
Mm.

Markus Rimmele - DigitalituM (11:33)
out of the mind of the senior service engineers into your platform? What's the magic without telling us the magic?

Sina Volkmann (11:46)
magic. Okay, I can roughly describe it. So first of all, as I have no technical background, I was always interested in the people and not in the machine itself. So I was wondering, is there any structure and is it worth to think of what is in these people's heads? What is not that's not documented? And then I came across and that was the first and it's also a very important point to understand. And we also try to educate our customers.

Markus Rimmele - DigitalituM (11:56)
Mm-hmm.

Mm-hmm.

Sina Volkmann (12:15)
towards is the difference between data and knowledge. And there is some very old literature on that and I have studied it like a month or so to understand that what is written down and what is some kind of information is not as valuable as someone who is able to interpret this written information. And that's the first difference. So first of all, it's about is it really everything even if it's structured?

all in the data that's relevant. And when it comes to very, I would say standardized processes, so let's say routine, some kind of maintenance work or so, that's easy because it's very easy to understand. You can repeat it and then it's in these people's heads. But when it comes to complex issues and most of the time it's identifying what the problem is, what I would say caused the breakdown, then it's always this trial and error thing. So

Markus Rimmele - DigitalituM (12:53)
Mm-hmm.

Sina Volkmann (13:11)
These youngsters, they call the expert and they ask the expert, okay, if I hear this and that noise and if the temperature increases to this and that point and pressure, I don't know, decreases, what might be the reason? What might be the root cause? And then the expert starts asking certain questions in a certain order. And this is based on experience only.

And most of the experts tell us that it's not written down in any troubleshooting chapter, in any documentation. And also gathering the machine data is not sufficient because you only would get some kind of temperature value. But whether it's too high, too low, or in the range is something the expert would need to add to this data anyway. And at that point, we understood that it's only about asking the right questions and get this valuable

built out of these people's heads. And this is also why we firstly did not involve any machine data documentation, to be honest. And it was just the reasons because these complex issues, they are in these people's heads. And it's exactly what you mentioned. If you have a look at the small companies, also the large ones, I guess they are even worse. It's not structured at all. They don't have any clue of which machinery is installed, where.

What is the customer, the user of the machine? Has there been any changes to the machinery that I would say somehow influenced the process? Maybe also was the reason for new failures coming up? So they don't really have an idea and it's not represented in the data. So to make it short, it's only asking the right questions and bringing the information into the right structure. This also means that we would need at some point talk to these experts.

Markus Rimmele - DigitalituM (14:59)
Mm-hmm.

Sina Volkmann (14:59)
And the magic, if you ask me, is not the AI we have developed, which is a specific one dedicated to that problem. The magic is how to spend, I would say, only a few hours with these experts and asking for the most important knowledge in the fastest way possible. And this requires some structure. And we have developed a structure to especially digitize

knowledge on troubleshooting because that's mostly the most painful part in service. And this structure is not only very efficient, but it's also effective because the AI we have developed just refers to the structure and makes the knowledge available at any time, any place, but in a correct way. And that's the additional downside of using data and documents and any AI.

and a generative AI, anything, for example, chat GPT or Microsoft Co-Pilot, because most of the time this AI, these solutions use or refer to is based on speech. This means that if you feed some kind of solution with documentation and the wrong AI, what you would get is some rough ideas on where some information is in and I know some general information on machinery, for example, on equipment.

But sometimes in most of the cases, you won't get any precise information on a specific root cause. And also you just have the problem of maybe the solution and the result is even wrong. This is not a problem if you have a look at generating texts or trying to create content, then you're going to rework it, right? But if you have a look at how a service needs to be done and you want to have it the first time right, and there's also some kind of

Markus Rimmele - DigitalituM (16:37)
Mm-hmm.

Sina Volkmann (16:53)
safety issues if you just recommend the wrong solution or the wrong group course, then it doesn't really help if you have, I would say, any information and any rough idea on what could be the reason. It needs to be very precise, it needs to be correct, and it needs to be, I would say, the concentrated knowledge of different disciplines put into one recommendation.

And then you have it and then you go and then you know, okay, that has been the root cause and how to fix it. And it's so complicated to think this whole process to its end. It's easy to put everything you know on a paper, okay, to just retire and ask some questions like, please, could you write down anything you've ever known? Okay, the expert starts writing. That's not difficult, just time consuming. But it's difficult to understand.

What is the right structure to do something with the knowledge, right? And this is the answer we give. So during this whole process of digitizing the knowledge, providing it in a guided way, and also making it somehow self-optimizing and updatable, if that's the word, right? Because knowledge of today might be not valid anymore tomorrow.

Markus Rimmele - DigitalituM (18:09)
Mm-hmm.

That sounds all very interesting. One comment on your point of senior technician writes everything down. And my experience writing things down is the topic what every person in field service and maintenance love and making very precise service reports and writing stories is what

everybody wants to do, unfortunately it's actually the opposite. So the quality what you normally get out of that is very low. That's why another approach and I think the approach what you guys do is they are the right way of asking the right questions to get answers and then my understanding is you put this answers into your

application. So guide us a little bit through that workflow. It starts with that interview with the senior technician. And you mentioned it takes about an hour, two hours. From there, that's the content creation, I could say. And what happens next?

Sina Volkmann (19:40)
Okay, so when we start a project, we firstly define what's the scope. What's the critical machine we talk about in your production, the cash cow you sell, the most service-intense machine. Let's talk about, for example, the drilling machine, okay, because it's just easy for everyone to understand. What we do is we ask for the two to three most experienced guys, some...

Markus Rimmele - DigitalituM (19:46)
Mm-hmm.

Mm-hmm.

Sina Volkmann (20:06)
mechanical with a mechanic background, some baby process engineers, some electricians. So three people. Bring them into one place. be on site, could be remote, like we do it. So it's both possible. It's just a way of moderation and structuring. And then we will start with a simple question. The question is, okay, what has been the last failure, the last service case you remember that was like

crucial, you spend a lot of time. They start with some kind of icebreaker. And that's also some kind of not only asking the right questions, but also somehow motivating and engaging the people. And then it will take about, as you mentioned correctly, one hour, one and a half hours. And we would continue asking questions like, okay, just remember that failure, that observation, what has been the root cause. And we're trying to always

understand what has been observations and symptoms and what have been root causes. And what we start building is a matrix. So some kind of clustering and differentiation between observations and root causes. And the magic is now that we combine both with probabilities. And as soon as you ask, for example, if you had some kind of wrong settings, how often did it lead to

Markus Rimmele - DigitalituM (21:21)
Mm-hmm.

Sina Volkmann (21:31)
this and that observation. Then you are in a position that you can use AI, because AI is nothing more than working with and changing probabilities. And this is something you wouldn't find in any documentation or so, because it's mostly on your table saying this is the symptom, this is the cause, this is the routine to fix it without any probabilities. And after one hour, you notice with these experts that they somehow get stuck.

So I would say that then 80 % of the most crucial failures it's talked about. it's, yeah, somehow written down. We directly digitize it in the app. So it's really like one of us is asking questions. The other one is putting down these keywords. And that makes it quite comfortable for those people who love writing because they don't love writing. They just hate it. They want to speak about it. They want to talk about it.

Markus Rimmele - DigitalituM (22:12)
Mm-hmm.

Yes.

Sina Volkmann (22:27)
They are very enthusiastic when they talk about the failures, but they just don't take this pen and paper. And we do it for them. For the keyboard. So not the one, not the other. So we do it for them. And when we digitize this first knowledge base, everything else, you call it the AI magic, happening in the software. So basically we don't use AI to get it out of the documentation because we...

Markus Rimmele - DigitalituM (22:34)
or the keyboard.

Sina Volkmann (22:57)
use our moderation techniques to get it out of these hats, but we use AI to transfer this knowledge base. For a drilling machine, that would be after one and a half hours, I would say 50 different observations and 40 root courses. So the most crucial one, and that's fine because you don't want, I don't know, you don't want to have them like for a week. They didn't have the time to answer your questions for a week. And I would say after these one and a half hours, they are, they are also exhausted.

They want to continue doing their work or the phone is ringing, okay, there's a service case, so I need to fix it and then we're done. And what happens then is we call some kind of managing director, some kind of other stakeholders that don't have the knowledge but are interested in the project, and also some kind of newcomer. So could be a younger service technician, could be also someone from a different industry but who was just hired to, I don't know, operate the machine or so.

and we simulate a failure. what we do is we ask, so, I don't know, Justin, just imagine you're in front of the machine. The first time you are just without an expert, it's only you, and you would have a problem. So do you have one in mind? And then there should be one use case. And what we simulate is our second part of the software, and it's in step-by-step assistance.

Markus Rimmele - DigitalituM (24:25)
Mm-hmm.

Sina Volkmann (24:25)
And what this assistance does is it retrieves the questions or the symptoms out of this metrics and ask them, yeah, it's some kind of query. So the user has asked the symptoms in a certain order. Can you hear a special noise? Is the machine up and running or is it just crashed? Anything, can you observe this? Can you observe that? And the youngster only needs to say, yes, no, skip, right?

Markus Rimmele - DigitalituM (24:53)
Mm-hmm.

Sina Volkmann (24:55)
we don't fantasize, so we don't really add anything what is not in that knowledge base. It's just that we go through this matrix and ask these symptoms in a very logical way and try to get to the most probable root cause the fastest way. And the advantage is that if you think of all these technical terms and if you ask a newcomer to search for something in a database, for example, and just in case the new

coming person does not know what to put in to the system, not the right keywords, no clue of where to start this troubleshooting path, then he or she wouldn't find anything. And that's the problem. This is why our concept is to not force the user to ask questions, but to guide him. So it would be like always one first symptom, one initial symptom we would ask.

And from there on, if the user says yes, no, or skip, he is guided through the system dynamically. And there will be a root cause we are going to recommend. And this could be, for example, settings of parameters, any mechanical issues. And we would say this is the most probable root cause for your problem. And this is the recommended routine of how to fix this issue.

And everything that's not in the software will not be added or with some other magical process be automatically mapped to this issue because we want to make sure that we stay correct. But what we offer is some feedback loop. This means that in case the assistance or the identified root cause was not correct, the user has the chance to feedback.

the correct root cause, maybe an additional symptom that he or she could observe, and also give some comments on, I didn't really get what you're meaning, so please put some pictures also. And that's the only way to take the last step, bring in knowledge from the field, because otherwise you would be always like setting up huge projects, starting from new.

Markus Rimmele - DigitalituM (27:05)
Mm-hmm.

Sina Volkmann (27:20)
starting for a drilling machine and you have another machine and another machine. And at some point you will notice that the knowledge is just outdated because there has been some changes to the machinery. And you will never come to that point that you say, okay, we are up to date. Yeah, and this is the last and the most important somehow task and also challenge we are facing because at the end some operator needs to put in again a new symptom, a new cause, feedback.

Markus Rimmele - DigitalituM (27:30)
Mm-hmm.

Sina Volkmann (27:50)
saying this was correct, this wasn't correct and feed the system with new content. And this is also something we provide. Exactly. And most of the companies, start with and they also focus on the problem how to get as much knowledge into a system as possible. That's not the question. Ask yourself, what is the structure and what is the way how the knowledge needs to be provided to whom and when and will it be correct?

Markus Rimmele - DigitalituM (27:55)
in order to keep it up to date.

Mm-hmm.

Sina Volkmann (28:20)
And starting to think of how to solve this problem from the end is the most effective way you can do it. And if you want to be like today correct and tomorrow correct, you need to have some self-optimizing logics in the background to somehow make this knowledge base flexible, dynamic, and also growing over time. That's how we do it. It works for every machine and no matter which one, the same way.

Markus Rimmele - DigitalituM (28:47)
That sounds pretty fascinating. Let me try to summarize that back up. So the one major problem is that when you have technical documents and if you search in technical documents for a problem, I always say your search is digital. You either find nothing because you used the wrong search key, like in a PDF.

Or if your search key is very general, you get 1 million hits and you don't find anything. And that's where with asking the question, put that in a matrix and getting onto your platform, kind of a different way of doing that. And the result is what I call this troubleshooting tree where you put a symptom out and the user has to tell yes, no,

Sina Volkmann (29:19)
If it's okay.

Markus Rimmele - DigitalituM (29:42)
or skip and by going through this question answer, on the other hand, the probability of a root cause is getting higher and higher, which is then most likely the problem which needs to be fixed. And my understanding is that you guys have then also a workflow in how to solve that particular problem with a step-by-step instruction, do this, do this, do that.

Sina Volkmann (30:10)
Thank you.

Markus Rimmele - DigitalituM (30:12)
Also, you guys don't kind of take 100 % of the knowledge out of a service technician. You focus on the major topics and then the rest comes over time with what I understand as kind of a self-learning, self-updating by adding additional symptoms and root causes and with this, the system stays up to date.

and actually grows over time,

Sina Volkmann (30:45)
Yeah. And that's just due to the fact. Yeah.

Markus Rimmele - DigitalituM (30:46)
because you don't want the static system, you want a self-learning system which over time is getting bigger and bigger. Did I sum that up correctly?

Sina Volkmann (30:59)
Exactly. And we have like these use cases where you have very old machinery and old industry. So we are in steel industry, we are in motors, how they produce like very traditional products, but we are also in new technologies and new way of producing or providing energy, for example. And these are all technologies where you don't have existing data. And there's

such a transformation going on that productions need to be more sustainable. They start building hydrogen-based technologies and productions. when you build up such new machinery, what is the initial point you start with? Some experts sitting in a room and thinking of these could be things or issues that could happen in the future. And we want to be prepared for that.

And this is why we decided to not stick to 100%. Like you don't have 100%. If you are the machine supplier, you would have 100 % of your technological know-how. But the operator of the machinery has the additional part of the knowledge. This is the process and the product knowledge. And you need to be open to involve this knowledge. And this is the only reason why we did it that way. And also because the algorithm is that, I would say, specialized.

Markus Rimmele - DigitalituM (32:04)
Mm-hmm.

Sina Volkmann (32:22)
also dedicated to the problem that still we would identify the root causes with a 90 % accuracy. You won't get to this accuracy with a chat GPT asking you chat GPT, okay, could you please tell me why the tolerances are incorrect? Or could you please tell me why my tool is, I don't know, breaking? These systems will tell you anything, but it will be quite generic, could be wrong, and not at that accuracy, and this is what we're gonna...

what we've gone out to you.

Markus Rimmele - DigitalituM (32:54)
Gotcha. One more question to that S on a production shop floor or from a machine builder point of view, they have a lot of different models and products of their machine. How do you kind of differentiate if we talk now about machine A or about machine B? Are these kind of two different data pools or is that all in one big

Pocket, can you explain that a little further?

Sina Volkmann (33:29)
Yeah, so.

The most important, I would say, challenge to overcome in industry is the interlink, interlinking of machines and systems, because it's a process, it's a production and it consists not of very isolated production steps, but they somehow communicate with each other. And what we noticed when we developed the solution was that we would need to somehow slice the production into different

technologies, this is different machines at the end. We will set up these knowledge bases for a certain scope of a machine, not for example a system, not a component, but it's a machine for example in the size of a drilling machine. Then there will be the next one, the next one and the next one. If it's the same technology, so if it's just a drilling machine of a different supplier, then we would copy and paste the content.

Sometimes it's 20%, you could reuse, sometimes it's 80%, depends on how many changes you have done to the machine. But it's quite a lot. So most of our customers, they start thinking that, okay, none of these machines is like the other. It's a very special purpose machine, quite complex. This is always the pre-assumption on when we start cooperating with them. And then they notice that when a customer or a youngster is calling, they are asking the same questions. So there is a way of copy and pasting it.

Markus Rimmele - DigitalituM (34:28)
Mm-hmm.

Sina Volkmann (34:57)
And we set up these knowledge bases for certain machines, we can bring them together. So we can interlink them because some of these issues, they occur, for example, at the very beginning, but you notice them at the very end. And you could notice them here or here or here. And that's how reality is. And this is why we decided to bring them all together. And it's also quite difficult model.

Markus Rimmele - DigitalituM (35:11)
Mm-hmm.

Sina Volkmann (35:23)
in the background we developed. it's not another AI, but some mathematical model on how to interlink systems. And the user then would have some symptoms, some of the observation, and would just again press yes, no, skip. And we would decide in the background which metrics or which knowledge base to go into, because it could be, again, several ones. And this is why our goal is to set up a knowledge base for the whole production.

because they are all somehow, if there's one breakdown of one machine, the whole process is somehow limited or even stopped. And this should always be like the highest goal, being available for the whole production 24-7, all the different shifts and all different languages, because it's so complex and you would never know on where in the production the bottleneck is and where it's originally caused.

that you at some point need to bring this all together. And this is how we do it. And to not, would say, somehow to use the resources of the experts carefully, we try to slice it at the very beginning and integrate automatically.

Markus Rimmele - DigitalituM (36:44)
So who is a typical user of your application? Is it a machine builder? Is it a factory with a production? Is it both? Can you dive deeper into that? Who are your clients?

Sina Volkmann (36:49)
you

Yeah, so when we first started with machine manufacturers, we noticed that they don't have enough service technicians to serve their installed base. Now it's 60 % machine operators, so global producing companies. Siemens is one of our manufacturing companies. Siemens is one of our largest customers because they noticed that although the machine manufacturers...

Markus Rimmele - DigitalituM (37:10)
Mm-hmm.

or manufacturing companies.

Sina Volkmann (37:32)
provided some digital solutions, the gap is that large that they started to transfer the knowledge internally and use their own knowledge to do their own maintenance work. And this means that our clients are all those companies doing service for industrial machines, meaning machine manufacturers with, I would say, 20...

Markus Rimmele - DigitalituM (37:45)
Mm-hmm.

Sina Volkmann (38:00)
10 to 30 % of their revenue done with service. are aware of that service as a business model and they want to expand it in the future. And those globally manufacturing companies that have, I would say, at least two or even five or 10 different production sites and need to transfer knowledge from one production to the other because if you just ask a person to...

an employee to travel. It doesn't really make sense, right, because it needs to be the perfect and all-knowing employee. And this is the other side of our customers. So those who do their maintenance independently of any machine supplier. And these are becoming more. And what's also interesting to recognize is that there are some service providers.

So I would say again, independent companies doing service for different machines of different suppliers because they notice that there's some kind of gap.

Markus Rimmele - DigitalituM (39:01)
Yeah. And let's talk now about the international side of things. When we talk internationally, we talk about a lot of different languages. There's German in Germany, there's English over here. If we go to Mexico, it's Spanish. And one of the issues is that the know-how of a service technician is mostly in one language only.

Sina Volkmann (39:07)
Yeah.

Markus Rimmele - DigitalituM (39:29)
So how do you scale that knowledge into other languages and how do you guys solve that particular problem?

Sina Volkmann (39:39)
Yeah, so we first have the problem that if you have a look at German only or English only, two different service technicians use a different word for the same thing. So it's not translating into different languages, it's translating from one term to the other. So how we do it is we're working with some kind of glossary in the background. So if customers or companies have some kind of...

a glossary on, this is the official term. This is the way our customers also name it. And this is the English, I don't know, Italian translation or so. Then we could easily build some APIs or interfaces and get this translation out of the glossary. Most of our customers don't have any standardized way of terms and how they are translated.

Markus Rimmele - DigitalituM (40:25)
Mm-hmm.

Sina Volkmann (40:32)
And this is where we decided to integrate some automatic translation into, would say it's 25 different languages automatically. So you only press a button, then you have the same word in English. The problem is that if it comes to technical terms, no open source translation module would be able to translate it correctly. I would assume, and this is also the feedback we get. So our customers have the chance to rework it.

Markus Rimmele - DigitalituM (40:40)
Mm-hmm.

Mm-hmm.

Sina Volkmann (41:02)
and make their translations. So they have 80 % translated by an automatic translation module. And then they rework the last 20 % and say, okay, that's not correctly. We call it this and that. And then this is the fixed translation in the system. Sometimes we also initiate some kind of standardization process when it comes to error codes or translation. So this is really a blank fields that need, they have so much optimization potential.

because companies start to understand that what service technician A or programmer A does and writes down could be relevant to person B. But person B is, I don't know, speaking only a different language than person A does and is of a different qualification background than person A is. And this is why we initiate so many cleanup work.

with our customers because they just noted, shit, we need to have some kind of glossary. And this is how we solve the problem. And I had a conference call this morning. So we had a MyStore meeting with the German machine manufacturer. We had two experts in the call and we were discussing exactly this issue because they have customers in, I would say, Africa or so. So it's agriculture, it's the agriculture area. Two people.

both experts, three different terms for the same thing. And I was like, this is something we would need to streamline before we're going to translate it.

Markus Rimmele - DigitalituM (42:36)
Yeah, that.

Yeah, that sounds about right to my experience as well that you call the same thing in five different things in German even and then if you translate that you get all the different variations out of it. That's quite a challenge but good that you guys solved that as well. In terms of business model when a machine builder is for example a client and builds up this knowledge platform.

Sina Volkmann (42:46)
Yeah

Yeah.

Markus Rimmele - DigitalituM (43:10)
Is there a way that they can share this know-how to their clients, the operators of their machinery, kind of like in a first level support thing? Is that technically possible? And do some of your clients execute it that way already?

Sina Volkmann (43:33)
Okay. Assuming that's a very proactive and innovative machine manufacturers, they do exactly that. So what they do is they use FindIQ as a white label solution. They bring in their knowledge and they share it with their customers. They put on their logo, their colors, their design, and they provide it as a self-service solution.

Markus Rimmele - DigitalituM (43:59)
Mm-hmm.

Sina Volkmann (44:02)
What they don't do is like they don't ask their customers to pay, I don't know, 50 euro per month to have access to a digital solution because it's just that most of their customers don't see that value in a digital solution that they see in a machine. It's just the perception of value nowadays. What they pay for some kind of service package. So it's a service level agreement and the machine manufacturers provide some

Markus Rimmele - DigitalituM (44:20)
Mm-hmm.

Sina Volkmann (44:30)
hours of their best experts, some availability hours, then they have some remote assistance, for example, they have some hotline or some chatbot, and then they have these expert system, what they call. So it's find a queue, not white labeled, right? So then it's the expert system or so. And then the customers pay for it. So it's part then of a service contract, I would say. And just exactly.

Markus Rimmele - DigitalituM (44:52)
Yeah, so it's a package with a lot of different things and find IQ can be a wide label solution in that whole package with the whole purpose and the value to the end client to reduce downtime. Because what they can solve by themselves relatively fast versus

Sina Volkmann (45:03)
Yeah. Yeah.

Yeah, exactly.

Markus Rimmele - DigitalituM (45:18)
calling somebody, waiting and then get an answer is helping to reduce downtime.

Sina Volkmann (45:24)
Yeah, and it's not black or white. So we won't be in a position also if we have like all the knowledge in our software. So it's not possible to cover every issue. That's not the goal. It's, okay, what is the issue the customer calls you like twice or even more times a day asking the same questions. And you would always like me to

Markus Rimmele - DigitalituM (45:37)
Mm-hmm.

Sina Volkmann (45:52)
answer, okay, have you like turned it off and on again or so? This is annoying, especially for those technicians who want to cope and solve more complex issues. And it's not that we reduce like, we shift all the cases from the machine supplier to the operator. It's just that everyone could solve the problem with their knowledge and their best capabilities.

Markus Rimmele - DigitalituM (45:55)
Mm-hmm.

Sina Volkmann (46:17)
And there will be more room and more time for proactive service, for some consultancy work, for production optimisation. So there's a lot to do more than reactive service. And we want to relieve these experts, also those in the development department. When they develop the machinery, some of them are called up and I'm always like, okay, is it the job you've been hired for?

You are here for developing new machinery and not answering stupid questions from the customer. So we want to relieve these experts from these issues to bring companies in a position to grow.

Markus Rimmele - DigitalituM (46:51)
Mm-hmm.

Yeah, that's actually a good point. I like that. Where do you see the future in all of that? Where does the journey go? Manufacturing, AI, different tools, things are shifting. How do you see the things going forward as trends, but also in challenges?

Sina Volkmann (47:06)
Yeah.

Mm-hmm.

I have. Yeah. So when we started, find that you had a clear picture on how the future of service could look like, but I'm obviously like not equipped with any special ability to do so or to predict the future. But at least what I can see in Germany and also with the few customers we already have in the US. So.

Markus Rimmele - DigitalituM (47:40)
Mm-hmm.

Sina Volkmann (47:57)
Basically, you have the operator and you have the machine supplier. And both have some kind of service they do, depending on how open and how service oriented the machine supplier is and how needed the operator is. Right? So service will go to these, I would say, those players who have the knowledge and some kind of human resources, because you would need to, the machinery don't, they don't repair.

themselves in the future. There might be some kind of prediction of some failures coming up, but as long as machinery changes, processes changes, products change, you can't predict everything. So this will be a fact that there will be downtime, there will be complex issues, you need to have some intelligence and some people fall to some kind of diagnose and then find the root cause. And there will be a gap between these two players.

And what I perceive is, I've studied a lot on platform business models and also disruptive business models and how they develop their network effects. And as soon as there's some kind of space between two players, there's a chance for platforms going in. And what I can see is that these platforms are some kind of independent service providers because they have a lot of people in place. Sometimes it's thousands or 10,000 of service technicians.

and they have the know-how of different machinery. So they would be able to do service for whole production, this full service thing, right? And the only issue they have, they don't have like knowledge for all of these machines and all of their employees had, but that's easy to solve the problem because there's fine IQ. But at the end, if you have the knowledge and if you have the people, you will be a service leader.

And for the end customer operating the machinery doesn't matter anymore if the service comes from this machine supplier, that machine supplier, they want to have production productivity high and they want to have resource efficiency high. And this is what I see. So there have been a lot of trends I've been focusing on before working with knowledge transfer. And I'm quite sure that talking about the machine itself and talking about the data and talking about technology only.

is not the next step of innovation because all these technologies are there. They are available. They are available to the whole society. So everybody could use, could use chativity. The only problem we have is hardware and infrastructure. Do we have the capacity to operate all these AI models that are very data consuming? And do we have the human? Do we have the culture? Do we have the processes in place? Do we have the organizations? Do we have the mindset?

Markus Rimmele - DigitalituM (50:37)
Mm-hmm.

Sina Volkmann (50:45)
to adapt these technologies to daily business. And this is what we are definitely missing in Germany massively. What could be a bit more innovative and more open-minded in the US, that's what I'm perceiving. But those countries and those players who will be the winners of the future, they understand that it's not talking about technology, it's about getting the people...

ready for these technologies and also understanding the differences between AI, between data, between knowledge and all these passwords going around because otherwise we're talking about the next trend without having implemented it and I guess time is not there to do this once again,

Markus Rimmele - DigitalituM (51:29)
Yeah, as digitalization overall is not an engineering topic or an IT topic, what a lot of people think, it impacts the entire organization. And I also would say you can break it down to the three things, people, process and technology. You have to do all of them at the same time. If you bring in new technology, you have to...

Sina Volkmann (51:50)
Exactly.

Markus Rimmele - DigitalituM (51:57)
guide the people with it and you have to have the processes in place in how to use that new technology in order to get at the end the value out of it because you want to either solve existing problems or dilemmas and or make things better, faster, more efficient and or adding more value.

your end client by providing for example additional information which helps your client to be better. That's to my understanding of digitalization what's it all about. Do you agree on that?

Sina Volkmann (52:41)
Absolutely. And as I'm not on the technical side in Find IQ, so I'm not the technical guy in our company. I'm the one bringing these experts together and focusing on change management and project management. And I've done so many pilots and proof of concepts and let's try this and let's put these augmented reality glasses on. I'm always like...

Markus Rimmele - DigitalituM (52:49)
Mm-hmm.

Sina Volkmann (53:07)
most of these technologies, it wasn't about the technology, it just was about how open-minded are the people and how intense has, for example, a team leader or someone in charge and responsibility has been with his or her people to adapt the technology to their daily business. not about throwing something, throwing up or...

an iPad to someone and asking for asking him or her to start and use the software like fully independently and then you are there and it's productive and it's working the next day and everybody's like okay I've been waiting for this application my whole life that's not reality but the reality also is the reality is that you need to have people

as some kind of first movers, key users, then there will be the first one standing up and saying, okay, I might like that solution. That's the best feedback we can get. Okay, this could be useful. And I'm like, okay, is it now that you like the software? Is it only could be useful, but that's the first step. And then there will be second person, second person confirms what the first one is saying. So there will be two people.

And as soon as you have one key user and one first mover and the fast follower, then there will be more and more more more coming. But that's, as you said, it's not about only the technology. It's really about people and working with them. And that's also something we offer. So we also offer change in project management and consulting somehow. And this is why we are also in cooperation, because we noticed that it's not about giving someone the software and feel free to use it.

Markus Rimmele - DigitalituM (55:01)
Exactly. It's the implementation part and making sure it works. Coming towards the end of our podcast episode, if people are interested in this topic and want to learn more, how does normally the process work with you guys? After

Sina Volkmann (55:27)
Mm-hmm.

Markus Rimmele - DigitalituM (55:27)
contacting you and telling, hey, we wanna learn more.

Do you guys do demos or?

Sina Volkmann (55:34)
Okay, so this topic could be a huge one, then it's overwhelming and then nobody starts like at all. Or this could be a very precise thing and this is what we're trying to do. So make it as easy and as fast as possible. And if there's someone interested in our solution, there will always be a trial phase. So no one like buys a contract or so from us for...

Markus Rimmele - DigitalituM (55:45)
Mm-hmm.

Mm-hmm.

Sina Volkmann (56:01)
two years, three years without knowing how the usability of the software is. That's fine. That's just the way it goes. And then they asked for mostly four months trial phase and it's paid once. So we don't like throw our software without any charging or so, because we know that if there is a commitment and the customer pays for a trial phase, then there's also at least one person in charge on customer side and there will be one person in charge on our side. And then the likelihood of a

positive result of the trial phase is higher. And we do it only for one machinery with two experts and one youngster, and then we test it. And we also offer some kind of support. So it's a weekly support, so very close to our customers and we get feedback and we are there. We're just trying to get as much knowledge into the software as possible. We're doing these initial workshops. And after these four months, there will be a decision on whether and how to continue. So.

Shelby wrote it out to more people, more machines, different business areas. And fortunately, all of the customers who decided to do this trial phase have also have stayed. That's how it is. So mostly they decide before that. If they understand the topic is relevant, then there will be a short demo and a short call with us. We directly talk about prices, internal costs.

Markus Rimmele - DigitalituM (57:13)
Wow, that's cool.

Sina Volkmann (57:27)
what this might lead to as a return. And from there on, they can be sure that if both sides commit to it, then there will be a continuation of the pilot phase to a contract. And I can tell you, it's also because we put a lot of effort in this customer treatment. And yeah, that's how it works.

Markus Rimmele - DigitalituM (57:31)
Mm-hmm.

Pretty good.

Very good. And Digitalitum is a partner of FindIQ. So if you're in the North American market, contact us. We are happy to demo and bring you in contact with Sina and her team to discuss any details about the problem you want to solve.

All right. Thanks Sina for joining the Digitalitum podcast. Lots of insights in what you guys do at Find IQ and let's stay connected.

Sina Volkmann (58:28)
Thank you, Marcus, for having us.

Markus Rimmele - DigitalituM (58:31)
You're welcome. Bye.

Sina Volkmann (58:33)
Bye!