Founder Spotlight: Matthias and Max of AiSight

In this Founder Spotlight, we talk with Matthias Auf der Mauer and Max von Duering, co-founders of AiSight.

AiSight helps world-class manufacturers better understand their machines and drastically reduce downtime by combining sophisticated hardware and sensors, AI, and a cloud-based software platform. Matthias, Max and the team are based in Berlin, Germany.

Max and Matthias talk about:

Much like the subject of our last Founder Spotlight (ALICE Technologies), AiSight is a great example of a company built on deep and unique technology and having an important impact on large industries.

Enjoy the conversation and let us know if you have any feedback!

(Subscribe to our newsletter and/or our YouTube channel to see future Founder Spotlights)

The full conversation transcript is below (lightly edited for clarity).

_______________________________________________________

David: Matthias and Max, thank you for joining us for our Founder Spotlight. We really appreciate your time and you're doing it from Berlin, which is where you're based and it's a little late in the day. So again, thank you. How about if we start with an intro from each of you. Matthias, you're the CEO of AiSight, Max, you're the CCO, Chief Commercial Officer.I'd love to hear a little bit of your background before AiSight and leading up to when you started the company.

Matthias: Yeah. First of all, hi David. Yeah. Thanks a lot for the invitation today. And we're very glad to be here.My personal background and also what led us to, to start AiSight: I'm originally from Bonn, a city more in the Western part of Germany, and ever since I've been pretty young I've always been quite interested, more in technology side of things and like physics and math. And this also led me to study mechanical engineering. Then I went to Switzerland, to ETH Zurich. And afterwards, during my masters, I kind of figured that I was not just interested more on the mechanical side of things, but also more and more on the electrical side of things, which then led me to go to Berkeley, to go into a sensor lab where we developed CMOS chips for different applications.

David: Berkeley here in the US, in California?

Matthias: Yes, correct. Yes, at UC Berkeley. And I spent some time in the Swarm Lab where we developed a CMOS chips for disposable HIV tests. And this really got me started on the whole IOT topic and also the whole sensor topic. And I kind of got into it, I got super impressed by this field, then during the time there I had the chance to meet a young team from Clarity. Clarity is a young startup in Berkeley. And they are developing air quality sensors and sensor networks. I really love the mission because they are looking to improve air quality in cities. And then I joined them to work more on the hardware side and to develop the air quality sensors from scratch. So these are PM 2.5 sensors, and then we also integrated them into the sensor nodes, deployed them in cities, built the entire IOT infrastructure, and also the platform on top in order to predict how air quality is developing in cities.

So, I mean, this was a really cool time and I learned a lot but at some point I came back to Germany and I wanted to stay in this sensor field and joined the Porsche Digital Lab in Berlin. And there my primary focus was on how to improve production lines and how to bring IOT into the production field. And what was really cool there is that I had the chance to speak to a lot of people in the production side to develop some prototypes ourselves in different fields, but I looked at how are predictive maintenance solutions being used and how can we integrate them at Porsche. And what strikes me the most is that these solutions out there had like three major problems.

First, they were super complex and super long to integrate. Then they were super expensive because usually needed quite a long time to integrate them. And also the technology, etc. is quite costly. And then the third point is that the ROI for the customer was so long that we, for example, also spent many, many months and it was quite difficult to get to actual results.

And this was the idea then to build a solution in the spirit, which exactly addresses these issues and which is actually super fast and easy to integrate, which is inexpensive and which brings immediate value. And this was the idea to start AiSight.

David: Right. And that takes us exactly to what AiSight is. So we'll get into that in a second. Max, how about quick background from you on your background leading up to the company.

Max: Yeah. So I first I grew up near Hamburg. In a really small town in Germany. It doesn't properly exist. It's like an ongoing joke, but that's what everybody is saying. And then my background is in business and economics. And I met Matthias basically back in my bachelor's already. Our universities were not too far from each other. He studied together with one of my childhood friends. So that's how we know each other since 2011. So 10 years ago. And so I got into the whole startup field in my bachelor's. I joined a VC firm in Berlin. And then moved more into a hands-on role in an e-commerce startup, more in Southeast Asia. So I was working at HappyFresh, doing international B2B sales. So mainly in Kuala Lumpur, Jakarta and in Bangkok. And that's how I discovered my passion for sales. I went back to Berlin to do a masters in this field, in international management at ESCP Europe, more specialized on sales. And, yeah, during that time Matthias came to me with this crazy idea for a sensor for predictive maintenance startup. And then we started right away.

David: The rest is history. Super interesting. So actually before we get into the company and the product, which is obviously important, let me ask you a question about… you are based in Berlin and Matthias, you've spent time in California and Max, you spent time in Southeast Asia. I was going to ask: given your experience, what do you think about starting and building a company in Germany, and Berlin specifically? What's been great about it? What are the challenges? It's seen as a startup hub for Europe, largely I think. Tell me a little bit about that. How hard is it to find people, hire people? That sort of thing.

Matthias: I mean, Berlin for us, both of us like actively decided to live in Berlin. As a city it's a lot of fun and it's also super interesting because there's a lot of things happening. There's been like, there's a huge startup scene, what it's quite known for. But I think also the whole thing that it's known for changed quite a bit in the last few years. So, Berlin used to be very famous for e-commerce companies and for FinTech companies. But I think especially in the recent years it changed a bit more and more towards software companies, but also hardware enabled software companies. And when we were here, we were thinking, okay, Berlin is known for its the more e-commerce type of company.

David: Like Rocket Internet and the like.

Matthias: Exactly. And if you look at the industry in Germany, which is mostly distributed in the Southern part or Western part of Germany, and you would say also from a sales perspective, Berlin is probably not super close to the customers, but we actively decided to start the company here, especially for, on one side for the talent perspective, because Berlin has a huge attraction for talent and especially for international people to come here. Because I think it's probably the, or it's for sure the most international city in Germany and this attracts a lot of people to come here. This is one part of it. Second part is the network, startup network actually is really amazing. And we also learned this during the last few years how many events are going on. How many investors are around, how easy it is to get in touch. And then the third part, we also learned that the industry around Berlin, of course not as big as southern Germany, but there's definitely some companies around and we also have quite some customers here in Berlin, and we really enjoy being here in Berlin. Maybe you also...

Max: Yeah, one addition is the government funding. So we got a first grant here, so all funded by the EU and the municipality of Berlin. And so there was infrastructure for young founders to provide office space for free to get like a monthly allowance, like a grant. So we could just focus on building the startup. This happens a lot, especially in the early days. And this is, this is what makes Berlin really attractive. And then if you compare it to Munich or Hamburg, I mean, just everyone is complaining about rent increases in Berlin, but still to live here is comparatively cheap. And then if you go ahead and compare it with San Francisco, I mean, it's, it's a very big difference in terms of price level, but also salaries.

So it's definitely a very creative city, a very attractive city. And so all of these factors made it very clear for us that this is the place to start the business. And then also in terms of starting a hardware company in Berlin it seems a lot of fun for us, because there's not that many. Well, we're not, we're not a hardware company. Right. But we have a hardware component and this seemed fun to us and it helped also because people who come from ETH Zurich, for example, who have more like a physics background they don't find that many other companies here with an actual physical product.

David: No, that's really great. So this leads us right into the company and the product. But it is interesting to note what you said, which is that Berlin to some degree, truthfully or not, was known as let's say, lighter tech e-commerce and things. But you guys are a perfect example of much deeper tech, harder things to do and an actual hardware component. Which is great, I'm sure for the ecosystem too. So let's talk about that. Mathias, you said how at Porsche you saw these issues with how industrial areas or factories were managed and the bad ROI and all this difficulty. So tell us what the AiSight product is and how it makes that better.

Matthias: Yeah. So as you mentioned when I was there, when I was talking to the people in production, they were complaining a lot about production breakdowns in general. And saying that machines often fail and in total they account for like 20% of all production costs. But then if you really look at the solutions that are already in this field, because vibration monitoring as a technology has been around for around 20 years, but the major hurdles were so far getting the solution up and running and getting them up and running in a scalable way. And to not have one sensor that works on one machine, but to really make it scalable. And this was the idea, to really make the product for predictive maintenance, very fast, very cheap and with a very quick ROI.

David: And this is for sophisticated manufacturers, right? These are not tiny little factories or old ones or something. You were working at Porsche. I'm assuming it's world-class manufacturers. Is that right?

Matthias: Correct. Working at Porsche, like I was in the production of Porsche, it's probably like one of the most modern, advanced manufacturing lines that anyone can imagine, because everything is just-in-time. You have this line of robots working together and every minute counts, every minute matters. And yeah, for example, a one hour break down costs around 2.5 million euros. So there is no chance that these machines can have a stop. So everything has to be done in order to prevent machine breakdowns.

And the idea is then to make it super easy. And this was the idea to build this product. I mean, to go a little bit more into detail about what we're doing...

This is actually what it is. We are building this, it's a very easy to use sensor kit, which very easily goes and clamps onto a production machine.

David: With a magnet, I think?

Matthias: Yes. So it's a magnet, which is here. The magnet, it goes onto the production machine and the production machine can be any kind of equipment like electric motors, pumps, compressors, industrial fans, injection molding machines, robots. So this is really the cool thing. It works for a lot of types of equipment. And then based on vibration, magnetic fields and temperature, we basically determine what is the current state of this machine. Is the machine running normal? Or are there any patterns in the vibration signal that are changing all the time and that show a sign of failure?

And then the typical failures that are happening, they have like a specific vibration pattern and this can be recognized again. And then based on the change of it, we can say, okay, your machine is going to fail. And within the next two months, your machine will fail. So we would recommend you to already reorder this bearing and then in a planned shutdown just very easily and smoothly exchange it, and then you don't have these unplanned breakdowns. Because unplanned breakdowns usually mean a lot of stress. They mean then if you don't have the replacement parts, like the breakdown doesn't take one hour to exchange, but like one week to get the replacement parts, then do it. So we are improving in the end, the maintenance processes and then the production processes, etc. And in long-term depends also to improve other processes like quality control and other ones.

David: This is super interesting. So let's go a little bit deeper into the technology first. We'll talk about the business in a second. If I understood correctly, you've got this hardware element. And a question that I have is, it sounded like, I'll assert, tell me if it's wrong or right: three types of sensors the vibration, you said electromagnetic fields and temperature, is that right? And then, and then you've got software, which I'm assuming is what makes sense of all of this. Tell me a little bit about the hardware, if that's right, those sensors, and then what the software stack looks like.

Matthias: Absolutely correct. For us, it was really important to put a lot of knowledge and effort into the design of the sensor kit because we really focus on using very high resolution and wide frequency sensors.

So they basically enable also a lot deeper analysis of the machine and also really being able to identify the root cause of a problem and this way to catch like the smallest deviation on it, even on a bearing. And this was the matter about the sensors and then we do the full end to end solution. So it's the sensors, it's the whole processing, the software behind it. And then also the dashboard interface for the customer. And so we send the data directly from the sensors over wifi to the production infrastructure, and then to the server. And then on the server, we do different kinds of models that are running in in real time. And that are doing on one side a physical based approach to understand specific vibration patterns. And on the other side a full AI approach to recognize anomalies, to understand on a specific level, what is the root cause of the issue and then also use the prediction model to predict the time to fail. And then on the other hand, we have the customer on our dashboard. In like an easy and very nice way to understand for them and make it really consumable. Because most of our customers, they're usually not the vibration experts that know how to interact with vibration and which failure looks like what. But what we do is we give them a way to understand if the machine is in a good state, I don't need to worry. Or if there's a problem to really interpret it for the customer and give them actionable insights, like what do I need to do? What is the best way of maintenance? How do I improve my production and make it so this is really the mission for us to make it easy, understandable, easy integratable, and like really easy to use in the end.

David: Got it. No, that's super interesting. Let me ask you about the business side and maybe this is for Max as well. Tell me about the sales process. My assumption going in, tell me if I'm wrong is, because we talk about these highly sophisticated world-class manufacturers, I'm assuming these are larger companies, maybe sort of slow moving. Tell me about the process in selling to them. What have been the challenges?What's worked well? Maybe I'm wrong. Maybe it's not, maybe they're not slow. Maybe they're not that large. So give me a sense for sales and how you've approached that.

Max: For the target customers and you're right, it's some larger customers who are maybe slow moving. We had some exceptions here, but just thinking about DHL for example. DHL Express with whom we work with, that was a very fast sales process. So if the customer mindset is there, they know what they want, they know what they need, it's very easy and straightforward. We also do work with a lot of small and medium sized companies. I mean, Germany is the perfect market for us here because there's a lot of manufacturing. There's a lot of industry and a lot of customers in this area. Then in terms of sales approach, we have a very clear defined sales process, with a lot of it coming through our network, Matthias' network, my network in the German industry. We did approach a lot of customers on fairs before Corona. Now we also use online fairs but also a lot goes through social selling, inbound marketing, the classic channels. So we invest quite a bit on the marketing side and all of this together, we are trying to create this pull strategy where we pull in customers, provide them information. And it led to a lot of traction because the product really fits from all the product specifications, the customer's needs very well. And so then we usually just work with warm leads. So customers who are interested in us we then have the sales team who's taking care of them, working them until they do an initial pilot. And then we have a customer success management who take care of the customer and move them from the pilot phase into a roll out phase. And the roll out phase depends on the size of the customer. So we have some that rolled out one factory first, then a few others until they use the solution on a global scale. Then we're talking tens of thousands of sensors, but we also have smaller customers where it's just, let's say in the few hundreds. And so, so this is how we typically approach our typical sales process.

David: Got it. No, that's super interesting. And I think that you alluded to the fact that Germany has a ton of these medium-sized manufacturers. But they're also highly sophisticated, right?

Max: We call them “hidden champions”. They're the global market leaders for very specialized manufactured goods. I think we recently talked to the world's largest market leader in manufacturing of needles for medical appliances, syringes and for the textile industry. So it's a very, very high degree of specialization, but they run their factories with a lot of efficiencies. So uptime is usually crucial for them. And yeah, these are our ideal customers. You find many of them within a radius of like maybe 20, 30, 40 kilometers to the south of Germany. It's crazy.

David: Yeah. I was going to say, Germany is one of the top three manufacturers in the world, I think, even though it's obviously a fraction of the size of a China or the US. One last question on the business side, at a very high level. When you think of these enterprise contracts and how they're doing, you mentioned, for example, they may expand to new factories, new plants. What are some of the key metrics you look at for any given customer? When you think about the health of any given customer or business, what are the key things you look at today?

Max: We have one key metric: it's number of sensors installed. That's the key metric that we're also pushing for because the more sensors we have, the better our networks, the better our software. So this is really our key metric that we look at.

David: Got it. Okay. Super interesting. One last question, Matthias for you on the tech side. Do you have to do customer specific training on these models so that it understands their machines? Or have you seen enough machines now where a machine is a machine and you can tell when it's going to fail no matter what?

Matthias: I think this is one of the key challenges of our product in general, to make it possible to go from one machine to the second and then to many machines. And I think this is also where we put in a lot of effort and development. What we do is we build these generalized models for a specific machine type, which is for example, an electric motor. And then what we do is we go to the customer, we install our sensors, and then we let it run for one or two days. And then with this initial training data, we do an initial set up, which is automated, which then gives an idea of that kind of equipment, and the surrounding condition. And then based on this, we do an initial training and we do an initial amount of audits and initial analysis, which already gives feedback on whether there are specific patterns, which look like a specific failure already. Because it might be possible that a motor that we're putting a sensor on is already...

David: Is already failing.

Matthias: Exactly. So if this is not the case, if there's no problems, it's set live, and then it's ready to run. So we use these generalized models, but we customize them in like one or two days, but this is pretty automated. And this way we make it possible to use the generalized approach on a specific equipment.

David: Got it. Oh, that's super interesting. That's great technology. So in the interest of time, I have one final question. You're still a relatively small company, but you've been very successful, fantastic product, great customers. Where do you see the company in three years? What's your goal? How do you envision AiSight three or so years from now?

Matthias: We definitely want to be the European market leader at this point for predictive maintenance. And we definitely want to have very large coverage in the German industry. For us, it's really important. I think our product has quite a customer stickiness. So this means once we’ve installed in the factory, customers usually really love the product. They see how processes change. And so our idea is to get into as many factories as possible and also on European level. And then also to start more on the internationalization side. To go into other countries and also be way more present. And then on the other side, what we also see is that besides the vibration monitoring, there's also additional sensors and also additional things that we can do in the factory. We also want to move more towards a factory platform. And then from the data that we're gathering, there's a lot more potential and things that we can do with it in future. And I think this is really where we we are heading to.

Max: And if you think, think back to our core KPI, three years from now, we'll be in the millions.

David: Millions of sensors?

Max: Yes. Yes.

David: Yeah. Great. Hopefully tens of millions very soon. Okay guys, this is fantastic. Super interesting, great technology, great product. Thank you for taking the time. I appreciate it.