Voice of FinTech®

5 of 318 episodes indexed
Back to Search - All Episodes

Predictive analytics for trading with Aisot's CEO and co-founder Stefan Klauser

by Rudolf Falat
September 7th 2021
00:31:53
Description

Stefan Klauser, co-founder and CEO of Aisot, spoke to More

voice of Fintech. Okay, yeah. Okay, welcome to voice of Fintech, a podcast mapping out the swiss and global Fintech scene connecting fintech enthusiasts with startups, incubators accelerators, business angels and VCS and incumbents interested in partnerships. Voice of intake will help you navigate the Fintech ecosystem here. You can listen to the startup, found the stories what investors and incumbents are looking for when dealing with startups and find out more about resources provided by incubators and accelerators. My name is Rudy followed and I'll be hosting this podcast. Yeah. Hello and welcome to Voice of Fintech today. We're joined by Stephan, who is the founder of assault or a I sought, we'll find out how to pronounce it correctly. And as you would expect, we're going to talk about, you know, a i machine learning, Right. And uh more specifically predictive analytics in trading, why is it needed, what does it do, how does that compare to algo trading or complemented or other solutions out there and uh you know how uh Stefan became a founder after working for the government but also working at it and at the th Right.

So very curious to find out about his journey and his thoughts on this space. Welcome Stefan, how are you today? Hey Rudy, thanks for having me today. I'm doing very good. I'm looking forward to interesting discussion, brilliant. So can you paint a picture, how did you get to what you do today? I mentioned that you, you know, you started with the th and other places and now you're running your own business. So how did that all come about? Yeah, that's a that's a good question. Right. Um I walked around with the dream of being an entrepreneur myself for quite a while. Um So as you said, there were some stations in between. So I started after my studies actually, I'm a political scientist um and I started to work and in public diplomacy and I had a station in the U.

S. At the Swiss Embassy in Washington, D. C. And I came back to Burn and I started to work for the State Secretary for Education, Research and Innovation and I was responsible for international innovation collaborations. Uh so I really helped in a innovative companies in Switzerland um and and researchers to to connect internationally and to start projects together. Uh and then I think it was in 2013 or so. Um I met with uh Sandra Tobler who is also a founder and Ceo in in the start up these days. Right. Uh and we kind of talked about collaborations and we decided that we that we want to do something for the uh then just kind of really being in the start of being founded, um do something for the for the Fintech companies in Switzerland.

So um there were some popping up around Zurich in Geneva and we said, okay, maybe it would be interesting to connect them to similar companies around London and new york and we organized a field trip for them. Um and that's also where we can first heard or talked about the Blockchain and what you can do with this fascinating new new technology. Um so we kind of were going on that trip and I also back then always thought, oh my God, I really like what I do here, but actually I would even prefer to be on the other side and do something myself and and started um so when I left, the government actually wanted to start my own company, but then I was in contact with the professor at 80 ha his name is dirk Helbing from computational social science and he said that he actually would like to work with me and I could help him kind of um streamline a bit his strategy and the research that he's doing.

Uh and I could do that part time and part time, start my own company. Right. And I thought why not? That's actually not not the worst idea. And then it was pretty clear pretty quickly that that this was a bit naive so there is, there is uh usually no such thing as part time working at 80 are right because that was quite a big workload and and and soon I realized I wouldn't have much time for for starting a company on the side. And then also we won we won actually a project application for any eu project where we wanted to build What we call the finance 4.0 system. Uh and the finance 4.0 system and basically was was distributed incentive Ization scheme for more sustainability. Um so that communities could issue tokens for sustainable or collaborative behaviour and it would be kind of fully free to train if people would want to train such a system.

Um and so I he asked me whether I could lead to that project, So actually I let that project um at 80 ha for three years. Um and this is also where I met my two co founders and you know, in TN um Yeah, and then uh with them, I finally then decided to start a company, brilliant and especially that you mentioned Sandra because she appeared on this podcast already in 2019 I think, and of course the founder of Future a at the moment. Right, And okay, so understood you early on, we're thinking about having your own business, but I'm always interested in motivations as well. Why have you decided to start your own company, be your own balls, being an entrepreneur. You know, you had some contacts with the world of government or diplomacy academia, you could have done then, the other thing. So why have you always thought about having your own company?

Yeah, again, I mean, I was working with lots of founders and the innovative companies and I I always kinda admired what they have been doing and I also for myself, I was always an entrepreneurial person throughout my life, I was always organizing new things, even um when we when I was working for the government, we set up things like the Swiss Swedish innovation initiative and and other things that we kind of found it while I was there. So I always had a little bit this entrepreneurial gene in me I would say and this virus or whatever you would call it. Um so more and more over time, I I really kind of became clear that I want to be a founder myself and want to try to build a company. Um and as you grow older you realize that you don't have time forever, right? So and then when I when I left the government I thought it's probably a good idea to do eat them. But then when I got offered a job at 80, I also thought maybe maybe it's not bad to come from a to B when you start your own company, right, and I can certainly meet talented people there, that that's really what happened then.

And so Nina and t and they were actually in at the same chair or the same professorship but they were working on a different project. But interestingly enough that we're doing research on novel machine learning algorithms that can better forecast a Bitcoin volatility. Um and they were invited to present that in Singapore at the International Conference of data mining. Um and there was actually quite some interest in the audience for for what they have been doing and back in the days and that those were mainly arbitrage traders, right, where the arbitrage opportunities were still still bigger in crypto. And so they were approached by them and people ask them, how can you show us how it's done? And as they came back and asked me in such Stefan, you you are the guy among us that knows most about entrepreneurship and then these kind of things and you always talk about having a company, we have been approached by this company and this company and what would you say?

I mean is this something that is worth looking into? And I said, yeah, sure, let's let's look into what's possible with this technology, what's the potential market, how could we scale that? And we did a little bit of research and and we came to the conclusion that yes, let's let's do it, let's start a company. Um and that that was actually back in 2019. So what is a sod, I mean what is the problem that you're trying to solve? And as I mentioned before, you label yourself as a company that provides predictive analytics for trading. So why do we need it? What for Yeah, so we we so it's called, I saw it right, we call it like like like the iphone, but but I sought even though it stands for ai systems and operations in trading, um so it can all could also be a, I saw it or whatever, but we think it's, it's better to call it.

I sought and I also produces next generation real time insights for fur traders and the asset managers to help them take better decisions. Um Why are we doing this? I mean, look, it's it's quite clear there is a growing class of people around the world that can save money. Um Interest rates are at historic lows, trading costs are also declining. Um So it's probably not the worst idea for for many people to invest that money. Right? But how do they do that? Most people don't really want to do it or also are not interested in investing the money themselves. So they need an asset manager, wealth manager certified investment advisor. And what what numbers show actually from research that is that most fund managers underperform um as compared to a benchmark, which could be an index or so. Right. Um And we asked ourselves why am and there's certainly different kinds of explanations and one is fees, but fees are decreasing as we said.

And then we also see that there is an increasing complexity in financial markets. So yeah, like real time information for example, is is kind of getting a bigger and bigger share on the total of information that is out there. So it's actually quite hard to have all the factors under control that could influence asset prices. That can be things like order book information, news flows social media effects um and so on and so on. Uh and with that we also see an increasing influence of social events recently that could either be a pandemic or or social trading such as Wall street bets or assets, even assets reacting to social media right? Everybody everybody knows what Elon musk can do with the Bitcoin price with a few tweets um So how can these asset managers have an overview of what's happening in real time and get kind of their conclusions from it.

This is exactly where I thought comes into play, right. We provide these insights to the asset managers or the traders we give them a statistical advantage for for a clearly defined time period and the clearly defined um use case that that they need. All right. So I mean how how does it work in detail um are you talking about like a news crawler or are you talking about uh you know getting insights from the older volumes or the markets volumes and prices or how does that work in a in a bit of more detail. Yeah. Okay. That answer is it depends, it depends on the use case itself. Right. And on the time arising we would use different information for short term predictions for example for a one minute Bitcoin directional price forecast.

It would need other information than for a three months volatility forecast for for foreign equities. Right. But the most important thing here is that we actually have the tools or the machine learning tools and technology to dynamically combine the different data from, from very mixed sources. Um so that we can listen to social media, we can take the news, we can look at the order book information where available and our algorithms will independently decide which factors are important at which point in time. Um and thus ensure that that we actually always give the most accurate information to the clients in the form of the insights and signals that we deliver it to them. I see and well how good is it today? I mean the old predictors analytic system do you back test it as well and compared to certain benchmarks.

So you can see clearly that if I've used this, I I will be a better as activists manager than I would have been without it. Yeah, certainly we we have our internal quality control as I as we could call it right. We we measure the performance of our forecast in real time, how accurate they are. But then also we we can at any point in time we compute the P and L um of very simple strategies that would be based on our forecasts for example. Um and and then um kind of all our all our algorithms are actually or are models are retrained every night. So in principle we can take in all the latest information on a daily basis. What we can do is give clients a statistical advantage about things like volatility volumes or even directions that the assets take, but it's not only about predicting something right these days because our products increasingly concentrate on longer term developments, meaning we take all available information into account and we apply latest research findings from machine learning on the steps that are taken along the whole investment process.

So we we also use machine learning for finding strategies or optimizing portfolios. In fact, our latest interest from clients is more and more around us producing volatility and co variance matrix and then applying these insights to a portfolio optimization process. Um and we also do more and more for classic as the classes for example equity portfolios. So we have clients that in principle have been doing their kind of portfolios in a classic way for for a long time maybe they didn't change anything in their process for the last 50 years. Right. It's often based on on fundamental insights that they have. Um and they increasingly get the feeling that they would like to get additional insights for their portfolio optimization process. Things like the cove occurrences are also our volatility forecasts for for certain asset classes or then they go one step further and then that's for example what we discussed with the turnkey asset man Turkey asset management program in the US, they would go as far as letting us kind of optimize full portfolio based on our machine learning insights and processes.

I see. So let's clarify the who are your target clients because you said before that people ideally not everywhere getting more wealthy and therefore the they have more money to save. Uh and uh on the other hand, when you talk about these solutions, it sounds like you target the institutional clients. So how does that work? Who are your target clients? Uh Yes, I'm sorry for the confusion, I mean, but in the short term we clearly our B two B oriented meaning the meaning the saver or the wealthy individual um with more indirectly profit from, from our services. We work with traders, we work with asset managers and we work with brokers in exchange depending on the product line that they would get from us and mostly also from the the horizon and that we look at with this product. So just some examples, right.

The broker that want to hatch the client flow there, we would give like some short term signals around directional moves or then for an asset manager has explained, we would rather do some, some things like weekly, up to up to bi monthly portfolio optimization with them so that it's all right. And how does your service relate to or differ from Algo trading? Right. Or can you embed it or can you work together or work against each other? In other words, you know, um you can also program some of these things depending on what happens on the market. Right? So are you going broader with with your service and and how could you use this together or against it? Usually we would be the ones that inform and the algo traders, right? So if you have for example traders that use algorithmic trading um for prop trading often they often have like their own signals and insights, right?

But there are often on the look for getting, getting new insights and new signals that give their their models new bias. Um This is where we would come into play many in many cases, we also see kind of algo traders that are rather active in high frequency or ultra high frequency trading. Uh and they're usually they have to trade a lot with which comes at a certain cost. Right? And the margins are rather low portrayed and what biden would give them is just um insights over a slightly longer time periods, let's say a few minutes. Um And with that information they can also try to kind of have traits that go into the right direction for a slightly longer holding periods than what they would usually do but would also kind of bring them higher, higher profits portrayed. So we work nicely together with algo traders in principle, especially the short term signal that they are not very useful for somebody that is still kind of trading by hand in that sense because you're most probably too slow to take up the information.

So what we do is we provide our signals through through an api um then then a trader can directly connect their algo trading systems to our signals um and and generate higher profits. Nevertheless, that's for the short term stuff again. Right? So um and then for the longer term Canada port for more portfolio related products and we, I mean the delivery methods doesn't play such a big role right there. Sometimes we even just send send a documentary in Excel or so giving the new weights of the portfolio so that there is the methods of connecting to our client is then much, much more low tech on that then. Okay, understood, understood. Now that leads me to the next question, how do you make money? We have different models for different client types. Um if you have a client that is in trading or in brokerage, we would usually go with the subscription fee to the signal and we pack revenue sharing component on top of that meaning that if we can more or less quantify what is the extra value that they generate with our signals.

We would usually like to take a percentage of that extra value. And for that we would then lower the subscription fee accordingly that keeps us gives us a nice upside. Right? So the more successful we are uh the more we would earn and also that kind of shows our clients that we that we like to have skin in the game that we trust our products and, and uh and that really helps approaching them in principle, when we, when we work with asset managers that are more fee based themselves, right? They usually take a percentage of the assets on the management as a fee. Um there we then just do like some fish sharing, right? And you're based in Zurich. So where where are you on your journey now in terms of products or geographic coverage and the ambition and my hint to you would be say something about global domination.

Right? Yeah, exactly. We we started in Zurich. We were also, we were also very committed to Zurich. Right? We have all been here at the data together. We all still live in Switzerland, the founders and one is in Geneva to two here in Zurich. Um and in principle we want to grow our team here. We want to base it here because we still also are very close 80-1 of our founders still still have support time position at the White House to just keep that relationship and be and exposure to the latest developments also in the field. Um and so we we actually like to be here and and the talent flow, that's that we have um around Zurich. Nevertheless, our our businesses is global. Right? So in principle, we said global first, we don't really cares that much where our client is from.

Um clearly there are some sometimes regulatory things to check before, before onboarding a client somewhere in the world, but in principle that's usually not a big issue. Um and we're a digital only company, we deliver all our services online, we can kind of on board our clients online, which makes things much easier for us. We nevertheless, this is this is just organically grown like that. Our clients so far, our client base is mostly in Switzerland. Um We're talking to some U. S. Clients right now and I guess we're getting closer to signing some of them um and then in principle, all all the markets where we have strong financial seen mainly mainly around UK, but also then in asia, around Singapore china um and clearly the U. S. Those are places that are very interesting for us, but it can also be any anywhere else where we have a client that is that is interested.

So we really, our our goal is to become the number one provider for real time analytics and and forecasts around the world. Great. So you kind of outline what's coming, but if there is anything else in terms of big milestones that you'd like to hit for the rest of the year or beyond, let us know. Yeah, so there's a few things right? We we went through the F-10 incubation program last year, um and that ended in March this year. We successfully graduated from that and we got got a convertible loan awarded for for being one of the top six companies in that program. Um And so we started to grow the team of it, and now we're actually at the point where we where we need to to really double down on that too, be able to move fast. I think the timing is great for what we're doing.

We have a good team. So now we're actually just about to to have our seed round. Um in fact, it's it's just tomorrow that I'll have another discussion with our lead investor. Um So um so we were actually raising raising around of of one million swiss francs to kind of kick starts that also on the, on the sales side to build a sale team and to make our deaf team bigger and stronger so that we can then also deliver if he scaled quickly and onboard new clients. Our goal is then to pass one million u. A. R. R over the next 12 months, I think that from from what we see today, that's still still very realistic. Um And then we would really like to scale quickly and and grow our revenues beyond beyond 30 millions over over the next three years.

So um really prepared to, to do that. Um And and it's going to be an exciting journey. Um We also have new offices from september. So we're we're moving and as the f the F 10 co working part is closed, they will close at the end of august the beginning of september. Um And that means all the companies here have to find a new place. We went together with some other interesting startups, like like Radio Alina um and Advisor and others, and we built our own Fintech floor that will start from september one at whole Struss. I'm in Zurich. Um So there will be, let's say, a new a new home for for interesting, innovative fintech startups in Zurich. Um And we're really thrilled to be part of that. And so yeah, I mean we're quite excited at the moment and and we hope we we can serve that way for for a long time.

Alright, well, great. Well, my last question is work and interested parties reach you. And who would you like to hear from most, whether that's potential clients or potential employees or investors? Yeah. Really? Uh Really? All of those. All right. Uh So where do they find you the most efficient way? Um Okay, the most efficient way is just to to write to me directly. My email is Stefan. That S T E F A M at I thought written a I S O T dot ch. Um I thought that ch is also our website where you can find more information. We're currently revamping that a little bit so that the new version should be online soon. Um Yeah, if you if you're a potential client, if your tray, if you're trading or if your wealth manager or an asset manager and you would like to get access to highly skilled machine learning team, please let us know.

We'll be happy to talk to you and see how we can support you with our products if you're an investor. Um Please also reach out to us even though our seat seat around it would be too late for that. But as as you all know, especially investors know after around this is before the next round. Um So please reach out to us as well and then clearly talents is the one thing that will be most important for for the success of ice. All right. We're a company that is not built on any kind of commodity or or anything or even a patent or anything. We need the best people to drive forward our innovative products. So mhm. We have good good access to talents coming from Mato and other places but certainly we don't see all of you out there.

So if you're interested to contribute around data science, machine learning but also from from the sales or the marketing side where we will need people as soon and if you like to be part of a dynamic startup and work with us, just contact us. Yes, we will be happy to meet you. Brilliant. Well thank you very much to fun and good luck to ice up. Thanks so much Rudy for the discussion and uh yeah, I hope to see you again soon. Yeah. Okay. Thank you for listening to voice of Fintech podcast if you haven't already check out. Also Voice of intake dot com where you will find all the episodes and additional resources related to the podcast. You can also subscribe to Voice of integer on apple podcasts, Spotify, google or any other podcast app that you like. If you have any suggestions on the topics or guests or how to make this podcast better for you. Please email us at info at Voice of intake dot com. Happy to hear from you. Thank you. Okay.

Predictive analytics for trading with Aisot's CEO and co-founder Stefan Klauser
Predictive analytics for trading with Aisot's CEO and co-founder Stefan Klauser
replay_10 forward_10
1.0x