I didn't have that luxury [raising a million for an idea], I just had to make something happen without any funding in the beginning. So that's why we took that path of repurposing an existing ecosystem into something new.
An interview with a Antti Karjalainen co-founder and the CEO at Robocorp, an open source stack for simplifying automation. Robocorp rasied $11.6M from Benchmark, Slow Ventures, firstminute Capital, and others.
Peter Zhegin:
Hello, and welcome, everyone, wherever you may be. My name is Peter Zhegin, and I'm hosting datafounders, a series of interviews with entrepreneurs and investors who work on data science startups. In this interview, it's number nine, I'm talking to Antti Karjalainen, who is a co-founder, the CEO of Robocorp, an open source robotic process automation platform. Pleasure to have you here, Antti.
Antti Karjalainen:
Nice to be here.
Peter Zhegin:
Awesome. So to kick off, can you briefly introduce yourself? Maybe tell us a bit more about your background?
Antti Karjalainen:
Yeah, sure. So I'm Antti, one of the founders, kind of the original founder behind Robocorp, and I'm the CEO of the company. My background starts out as a software engineer. I'm currently in Finland. But the company actually is in the Bay Area in the US. So we're split across multiple countries operating as a fully distributed organisation.

My background personally is is from Finland. And I've been studying software engineering, particularly. And then working as a software engineer for some time, and then gradually got into more entrepreneurial topics, went to Business School at some point, because we do have free education in Finland. That was an interesting weekend project, I'd say. And then through that experience, got more and more sucked into the world of business and eventually ended up starting a consulting firm, which then was later acquired. And after that started this company, that's now Robocorp.
Peter Zhegin:
That's a very interesting background. And as far as I can see, for you, it was more or less smooth transition. Right? So you started as an engineer, but you did have some interesting background, some interesting experience, unrelated to engineering. So I'm curious to learn what are maybe skills or learnings outside of the software engineering that you got across these different experiences. And that actually became very useful when you became a CEO in a startup?
Antti Karjalainen:
Yeah, so I think, from my day to day job, I use a lot of my engineering skills. I'm involved in product decisions, make decisions on various technical framework. So it's highly technical on my side. I don't need to go into the like, really the fine detail, but I do like to understand them on the high level.

Other skills that I have learned and I use like day to day, they are mostly on the business side, but it's not like your traditional business school learning. I do read balance sheets and these kind of things, and when we do funding rounds, it's good to understand those topics. But it is more of a general understanding of the world through a business lens where you can, through experience, you can apply that knowledge. And obviously, it's something that I'm constantly learning, at the same time, reading a lot of material, reading other books and so forth. But just having those general frameworks of thought I'd say the engineering framework, and then the business framework, and combining those can be a useful thing.
Peter Zhegin:
And if we look a bit deeper at what you call 'business frameworks', what would be the most important things that somebody with a technical background needs to acquire, before launching a startup? Is it marketing? Is it PR, HR, sales? Because there are a lot of different fields. But what are the most important ones in your opinion?
Antti Karjalainen:
Ah, that's that's a tough question. There's so many things to know, you listed bunch of them. I really can't pinpoint any, any of those. Everyone has their strengths, obviously, and you need to be aware of what's involved in, say, marketing and not to just overlook that topic, you need to understand what you don't really know, and make sure that you have people who know better than you. So marketing for me at least is an area where I'm constantly learning right now we go into market and actually hiring people for different types of marketing roles, and I'm constantly learning new things about them.

And so you just need to understand like what's around you? I'd say that just see the world outside of engineering, and realise that it's a big world. And it's a nuanced world as well.
Finding co-founders relevant to the business you are in
Peter Zhegin:
Yeah, engineering itself is quite a galaxy. At least what I can see that usually, a founding team tries to distribute all these different knowledge across the team, somebody knows a bit of marketing, somebody else, a bit of HR, etc. So what was your experience with finding co-founders? How did you actually meet these people with whom you started Robocorp?
Antti Karjalainen:
Yeah, so I picked my first co founder, Teppo, when I had the initial idea, I felt strongly that we need to start pushing for this. It was just like, distinct moment, I remember, I was driving my car. And I felt strongly about it. And I wanted to start something. So I enlisted Teppo, who was a business partner of mine from from my consulting days, he was actually the CEO of our consulting firm. And I just knew that we kind of hit it off pretty well, because we've been working already together.

So I think I think the criteria was that when we were working together, in the previous firm, we never fought together. We might have argued, but we didn't, didn't fight per se. So that's a good trade in a co founder. A lot of startups actually fall in pieces, when the founding team starts fighting with each other. So try to avoid that, at least to make your personality somehow match.

And then then the second co-founder that we had was, was Jouko who actually had a lot of experience on operating companies internationally. And that's what we wanted to do from the start. So we were a Finish team, and we wanted to start a company in San Francisco. And there's 10 hours of time difference and, you know, a long way to San Francisco from Finland, and we didn't have any network. So we wanted to get somebody in, not like a network person, I I'd advise against getting like a network person to your founding team, but at least somebody who has the general understanding, like, how do you set up your structure? And and how do you operate the whole group of companies that you're setting up. Not to say that anyone should set it up the same way as we did. But for us it makes sense to have somebody with that sort of a business side experience.

And then the fourth co-founder I enlisted was Sampo, one of the smartest engineers that I know and really hard worker. So we just wanted to have the raw engineering talent as well represented in the founding team. A technical founding team that we have mostly.
Peter Zhegin:
Did you meet Sampo as well through the consultancy?
Antti Karjalainen:
Yeah, actually, I had been working with Sampo before, he was my boss before. He was always giving me a hard time about my pull requests. So I thought that he is going to do well.
Ideation
Peter Zhegin:
This is an amazing story. And you mentioned that you were drive in a car when that idea, like hit you. Was it more like an insight or really it was an accumulating process, that you were thinking about this idea for a while, and in a certain moment quantity became quality?
Antti Karjalainen:
So the originally, it's a long series of consequences in a way. So I was at my University's Alumni conference in 2016 (the company started 2019). So 2016, I was at this alumni conference, and I bumped into this thing - somebody was presenting about RPA. And I looked at it - that's strange. That seemed pretty odd thing to do, automate applications in that way. And how come is that any different from software test automation?

I happen to have a pretty intimate relationship with software test automation, because a year before that, we had actually started a non-profit organisation around an open source project called Robot Framework and that's a project that has been around for more than a decade and it was a project where we knew all the all the key contributors. So we wanted to push that project forward, we have a great community. So I was involved in creating that first non-profit organisation around it. And we were planning to host conferences, and we've done that since. I had this kind of affiliation with this open source community around Robot Framework. And I thought - wait a second, that RPA thing looks exactly the same on the technical level as Robot Framework. How, is it any different?

Then I started to get to know RPA better, I started interviewing people, because it was really like, enterprisy, you couldn't actually download anything online or get access to any tools. So I just went called people and started interviewing them and understanding the space. And, and through that process, in the next year, our consultancy was acquired by a larger firm

While we were sort of fusing with that that bigger firm, I learned that they had actually people doing RPA internally. So when I first got my keys to the new office, it was a 700 person company, so it's quite large place and I got my keys to the office, the first thing I did, I went to the RPA department, they started asking around. Through that discovery process, from the first learning that and kind of making the similarity, understanding between this and open source world, and then then I had this chance to actually go really deep through this acquisition that happened at the same time. And I had access to all these people and technology, I started to kind of realise that there is actually a need for this, that we could build it. We just had our first Robot Framework conference, and I just felt that the community was so kind of lively, and engaged. I felt really strongly that, you know, I'm probably the only one person around who could make these all these connections and pull them together. So I just felt that if I if I don't do it, I'm wasting a huge opportunity.
Peter Zhegin:
Interesting, you touched upon multiple very interesting themes. Theme number one open source itself. The second theme is the idea of repurposing, maybe when you look at certain piece of technology, and you feel this unique perspective that allows you to use it somewhere else with very interesting consequences. Do you have maybe a framework or any reflections about this moment of repurposing, and how somebody can spot something that can be reused in a very interesting way?
Antti Karjalainen:
Yeah, good question. I don't know if there's a genetic framework around that. Because I had been involved in software test automation world for a long time. And I kind of understood how that works. I think there's a huge coincidence that I was on that lecture, seeing those people talking about RPA. Otherwise, I probably would have been like two years late to the game already. And so it was like really early realisation that thing looks like that thing. And we could do that better with Python instead of using legacy Windows tech. Unfortunately, I don't have any, like really generic insight on that. But the way way I thought about it was that now we could build something from scratch for RPA. Or we could take kind of an ecosystem that perhaps has been built over a decade. And use all the integrations, all the tooling, all the pieces that are available, just repurpose to tiny bit to to make it fit into this RPA use case. And that allows us to leapfrog at least like two three years of development effort. So we took that path rather than trying to build something from scratch. Maybe if I would have been in in, let's say, living in San Francisco, and I could have called my Stanford roommate buddy to, you know, invest a million in my idea, I could have taken another path. I didn't have that luxury, I just had to make something happen without any funding in the beginning. So that's why we took that path of repurposing an existing ecosystem into something new.
Peter Zhegin:
This is an interesting perspective. And I've heard about two approaches more or less generic. So one approach to repurposing is very deep domain expertise. So what I'm saying is that when you know something really deeply, let's say in your case, it was testing software and testing frameworks, then you're able to do analogies, and to see how to use that thing that you deeply know to something else. And another approach is really very contradictory to the first one is about knowing at the high level, lots of different things. So when you know, lots of different things at the high level, you can probably combine kind of a puzzle, that's typically what venture capital investors do. I know a lot of things, but my depth is, to be honest, not super deep...
Antti Karjalainen:
If you have that framework, and I think that fits to the first one. I often talk to people who are not deep, in the RPA world, for instance, and they tried to draw these all these kind of parallels. And to me, they just, like, somehow sound a bit silly, even, like, let's say, some of these legacy RPA tools, they might have executives who come and explain that, you know, this RPA tooling will replace all the test automation tooling in the world... Now it is definitely not gonna do that, for sure. And there's really, really good resource why it's not gonna do. But they are looking at it from high level, for them it makes a tonne of sense. But for me as a domain expert in that technology, it just doesn't begin to make sense. I think that first framework is maybe more useful.
Peter Zhegin:
Interesting. Interesting to know that for you, it works because I've got different options, different scenarios, and I guess there is no kind of one fits all solution.
Antti Karjalainen:
Yeah, depends on the thing that you're coming up with. in my case, it's highly technical.
Applying Open Source to X
Peter Zhegin:
Let's chat a bit about open source. Why actually open source? There are some other products that are not open source. What was your logic? And what's your framework to decide, actually? Can that problem be solved through open source or not?
Antti Karjalainen:
I think it is really natural to me to apply an open source approach. Firstly, because I have a developer background, so. So if you look at developers, for the past 30 years, the most significant improvement in software productivity has been driven by open source software. So that's just natural, it's like, you cannot build anything without open source. When we are building something like an RPA platform, it's a horizontal thing. It requires you to connect into 10s and 10s, and even hundreds of different technologies. And, you know, if, I would have unlimited budget, I might try to go and build it out for myself. But that would be still a world full of all sorts of cool open source projects that you can just leverage and combine it in in ways that makes sense in this application area. So it's just like a pure pragmatic developer thesis to go that way.

The second thing was that when I started understanding RPA, and then the world of RPA talks about citizen developers. And so they have this pitch that this RPA software is so easy to use that anyone can use it even your accountant could take an RPA suite and automate their own task workflows. But when I started going really deep into RPA, I understood that that's not ever going to happen. And really, I saw this kind of promise of this low code or no code tools that you can just drag and drop some boxes and combine them. And you all of a sudden make these really intricate automation workflows happen. That wasn't working at all. So whenever you would go into anything more complex than your trivial use case, it will break down and you will have edge cases and it's just good to make it work. Second was that non-technical people, business people were really not that interested in building automation in the first place. It wasn't their job to do that. So I understood that RPA development is fundamentally a space for actual developers or people with scripting knowledge, people with, you know, IT backgrounds. And so for these people open source makes more sense than proprietary tooling. So it was maybe a combination of you know, you have a horizontal product, and that you have a user base, that's technical, and they appreciate open source software.
OSS product management and monetisation
Peter Zhegin:
Yes, it's, these are both very interesting lenses to look at how opens source is applicable to something. I wanted to touch on product management, because typically, even in the normal startup it is difficult to prioritise. You have different customers, especially at the early stage when customers may push you to customise, but how do you prioritise what to build within the open source?
Antti Karjalainen:
That's a good question. In open source, you actually have this benefit that you can leverage the community as well, you don't need to build everything in house. We actually done a number of projects where we incorporate some sort of a key technology into our automation stack by working closely with the community. It might be work with somebody who is a co-maintainer of a project. And then just through funding or support, sponsor some feature development, we might actually contribute to upstream projects ourselves. And then we might fork a downstream project from another thing that we see useful, and use it.

When it comes down to like, managing, really a core piece of open source infrastructure project, and if you want to pull it to another direction, and the project lead wants to go to another direction, then that's a tough situation, and that becomes political. You need to be a bit of a strategist to to really play the open source, world the right way. Again, me being a software engineers kind of helps, because I understand what's happening and I can make the company level decisions on those open source tool matters.

We just like a few weeks ago, decided to open source a key part of our tool chain. And it's such a system that you really cannot take back after it's done, it's there, it's out, it's open, it's Apache licence, you know, everyone can use it. So you really need to be confident in what you do and understand the consequences as when you're doing those, decisions. But coming back to the product management side, in the end, when you are building your core offering it is the same kind of thing as in any any product management decision. You have some of those open source factors affecting the world around you, but you still need to decide who are you building the product for, what's your vision for the for the next three months in the next six months, and so forth. So it doesn't change it too much, but you need to be mindful of the surroundings.
Peter Zhegin:
It's very interesting that you highlighted that in open source, you can't turn things back. And another interesting scene with an open source is obviously pricing. What would you suggest? What would be your advice about pricing to anyone who tries to build OSS?
Antti Karjalainen:
That's a good one. Obviously, you need a working business model, just start something. There are projects that start by just being a cool piece of software attracting a lot of GitHub stars... but I would hope that you'd have some sort of an idea of how you're going to monetize that in the end.

We, for instance, use an open core model. What we have actually in the open source side is that we have this open source automation stack that's built on Python, the bottom layer of the stack is Python. Then we have robot framework, and a lot of automation libraries on top of that, that you can use to integrate different kinds of applications. Then we have a configuration format on top of that, that defines that software, that sort of that software robot, as we call it, that piece of automation. And then we have a tool chain on top of that to put it all together and make it a sort of a nice automation project that you can manipulate and configure and run it anywhere. So that's the really the open source side.

Then on top of that, well, obviously, we do support developer tools as well, that's just another thing. But on top of that is the operational layer that you will need to, to really use that automation stack in your kind of enterprise company setting. And we realised early on when we started to apply Robo Framework into RPA that, sure, it can automate things that you can run locally. But when you want to actually sort of operate that automation in a company setting, you need scheduling you need to be able to run it remotely on local networks, and so forth. So we started applying, like CI and CD tooling, a bit basic stuff like Jenkins on trying to see that could we actually operate this with CI tools. And it turned out, that was a huge hassle. And it wasn't really meant for that. So we figured out that, hey, we should create a platform that can be used to run this automation code anywhere in the cloud, inside containers, inside virtual machines, on local ,on prem machines, and anywhere that you need to run when you're running that in a company setting. So that's the proprietary crust that we built on top of that open core of the automation stack.

That's a pretty popular model right now, works well for many companies. For infrastructure companies, you have the basic hosting model, so you have like, database software, like MongoDB, or Timescale DB, you obviously have the hosting. There you have, you know, the scary proposition that Amazon is gonna copy you, and just host your open source database for you. And that's really scary. So I see that those companies are starting to apply a new licencing model. So it's open source, but not open source for for hosting. So they kind of are blocking those cloud vendors from hosting it for them. So that's obviously if you have that kind of infrastructure platform thing, that's, that's a good way to go. But be mindful of the, you know, big three cloud players competing against you. That would not be a nice feeling, probably to have.

And then the kind of the, in my view, the legacy open source model, which is, you know, Red Hat, you know, selling support and services. And I think that this false understanding in the world that open source equals to support and services. And that's just really not the case. And they are much more exciting business models, and pricing models and opportunities to actually monetize open source software. And that's mainly because of kind of vast cloud adoption in the world.
Learning through consulting
Peter Zhegin:
Thank you for highlighting all these different approaches. And very interesting to learn maybe how did you arrive to the model that you actually use? How the process of customer development worked for you?. Didi you talk with enterprises or with developers with the community to understand what exactly you can monetize?
Antti Karjalainen:
Thank you for highlighting all these different approaches. And very interesting to learn maybe how did you arrive to the model that you actually use? How the process of customer development worked for you?. Didi you talk with enterprises or with developers with the community to understand what exactly you can monetize?
Peter Zhegin:
That's interesting that you mentioned, this consulting component, because what I hear frequently among investors, and sometimes among data scientists and engineers, they kind of neglect consulting a bit. We as investors also neglected a bit like - consulting is scary, you need to stop with consulting as soon as possible and do product or platform or whatever, you can turn into recurring revenue. But what you're saying is that actually, there is a value within the consulting for you as an entrepreneur, right?
Antti Karjalainen:
It's not a typical bet, I'd say. But, you know, since I've been already in the software consulting world running a business, I knew how the space works. And I could use the situation that I had to my benefit of actually like trying to run a business on obvious services. So now with Robocorp a lot of our users actually, this smaller RPA shops that offer Robocorp as a service to their clients. So they are automating mid-market companies, small and medium companies, business processes with Robocorp. And our pricing model fits really well with them, our deployment model fits really well with them. But that's not a coincidence, because I had the personal experience of trying to see how that will work. So kind of stepping into the shoes of one of our customer groups, and also talking to those buyers before we created the company.

I know where that sort of attitude towards consulting comes from, and I don't say that anyone should go to consulting if they can avoid it. But for me, it was a learning experience and something that I could actually take and leverage into a product business.

But the model that that often is mixed is that - 'you know, yeah, I'm going to start a consulting company, I'm going to do some gigs for my clients, and then I'm going to build a product on the side'. And that usually doesn't work. So that's not the right way to approach it. Because then you are working most of your time to client projects, building some products really on the side. And then when the first client wants to adopt it, you just tailor it to their need. And you kind of throw the vision in the trash bin and off it goes.
Learning through consulting
Peter Zhegin:
Yeah, that's probably not the scenario that I would also suggest to people. What's important, in my opinion, is to to appreciate your experience. I wanted to look at also on timing. You were really thinking about that idea for quite a while. Were there any external events, maybe in the market, the economy, whatever, that provoked you to start it at the exact moment. Or were it mostly internal factors that came together, and you made a decision?
Antti Karjalainen:
If I could have started a year or two earlier, I would have done it. That time, I think the market would have been probably not ready for it. Right now I see a new RPA tool comes up every week, literally, like, is becoming a hugely crowded market. But they're all sort of replicating what the big three. You have big three - UiPath, Automation Anywhere and Blue Prism. And these are the really the top leading players in the in the space, there's sort of a group of second tier companies that have been around for a long time. And then there are all these upstarts that are popping up and replicating some aspects of the big three. For me, those upstarts are really not that scary as a competition. Because I typically check that, you know, if they don't have pricing available on their site, if I can't download the tools immediately, then there is no competition, because they're not really product driven companies at all, as we are.

Timing wise... when we decided to start, RPA had been already declared to be one of the fastest growing segments in enterprise software. And that was a signal that you know, there's going to be a big market behind this, we knew all the faults that technology has. But still, starting a companies you have a lot of inertia before you get to a state where you actually are credible offering so right now, we have a good product out in the market, and we are iterating on it fast, and we are getting traction from users and even larger companies are looking at it seriously.

But if I would have been able to be in this situation a year earlier, it would have been, you know, obviously, better. But you know, I think like some people say that the right, right timing is when it feels too early. And building something completely new when we started out, nobody had heard of applying open source into RPA. After we were already going, I saw that, you know, at the same time, there were a few other people starting similar projects, none of them really got off the ground properly. But we managed to do it. And we really are a kind of the dominating players in this segment of the market.
Peter Zhegin:
And how to think about the trends in RPA. Right now, more specifically, do believe that there are some space, some opportunities for other startups? So people who will be listening us, do they need to think about startups in RPA? Or it's like it's done, it's too late?
Antti Karjalainen:
Nothing isn't never done in that regard. And I believe in tha if you don't have any competition, that's scary, because it's, you know, you're on a market that hasn't been validated. So I'd rather go into a crowded market then one that doesn't have any competition, although it sounds kind of counter intuitive.

If I would have to start making money off of it right now, I'd start a services company and use Robocorp. Just so one of our partners just last week, they signed a contract for $500,000 to automate processes for legal services company. So if I would need to make money immediately I'd do that.

But the way we are looking at the market right now is that the mid-market, anything below large, large enterprises is fairly unserved. The market penetration is still fairly low, even though some of these players might claim otherwise. I think the market is still in an expansion phase where we kind of see a kind of wave of adoption continuing for the next few years. And the mid-market, like I said, it's untouched. So there's a lot of opportunity, so the need to automate is really prevalent across any size of any type of company. When you look at the world right now, on the agenda of any sort of CEO or CIO, running an operation, they'll have, you know, analytics, AI, and automation. Those are all on their agenda. And that's gonna be a huge thing as we see this market growing and adopting new technologies. Definitely not late to be in the game. But you know, there's some inertia getting started, obviously, always.
Timing, RPA trends and opportunities for startup
Peter Zhegin:
Correct me if my analogy will be bad. But let's, let's think about RPA and another kind of, hot, technology, let's say machine learning, right? How I think sometimes about the market - there are machine learning platforms like H2O, very generic platforms, right. And at the same time, a new wave of companies emerge, they do, for instance, synthetic data, or let's think about security in the context of models, how we can make them really secure, etc, etc, etc. So you kind of see the fragmentation of the stack Do you think that similar things will happen in RP and opportunities there will be more or less significant in terms of the sizes?
Antti Karjalainen:
There is already something like that happening to a degree. You have a basic RPA, which is like essentially automating things that you can do with a browser and desktop applications, and so forth. And we don't even talk about RPA that much inside Robocorp, we talk about automation, because we see that we are Python based stack, so we can apply to RPA, to what is called Intelligent Automation, which is essentially combining machine learning models and anything that can be vaguely resemble AI, to automation. And then you know, IT automation, and then just general, like enterprise software integrations. All of these can be put under the umbrella of automation that we focused on.

So but but there we have, some of the big players are starting to talk about really terms like hyper automation, which essentially means combining RPA to Intelligent Automation to process discovery, really these hype acronyms... But I'm seeing that the process discovery space is going to be one area that's going to be sort of separated from the main automation spec. So that's really about understanding what you can and what you need to automate in the first place. There's companies like Celonis is a big one. And then they are like upstarts trying to come in the space. That's like a crazy difficult problem to actually understand from the way people use computers and ways data flows inside your enterprise information systems. And looking at that, and seeing like, - hey, we could automate that. It's actually a super intimidating problem to be, but still, there's a lot of a lot of people trying to do that.

We see some things around portability of automation about, obviously, security of automation. This is a topic and then maintainability and, debugging and all of that. But I think with us, we are coming from this Python world, we can already apply a lot of software industry best practices. So we don't need to reinvent the wheel for automation, specifically, but yet this test attempts to form these categories, I'd say.
Peter Zhegin:
That's really fantastic. You already described a bunch of themes within the universe of automation. And suggest anyone who thinks about a startup to look at these potential sectors, and look at what Robocorp is doing, super interesting. Now, we covered a lot of interesting things, I'm sure. And obviously, we can cover a lot of more. But is there any question or theme, or topic, that we didn't cover, but you believe deserves to be covered?
Antti Karjalainen:
I'd say that for for that data crowd - don't overlook the automation trend. Because a lot of a lot of things that we talk about inside our company - that we are building essentially hands, you know, that can access a lot of places that can make things happen in your IP systems or in the real world. But really, the brains are going to be the decision making systems that the ML/AI space is building. So we see that there's a huge synergy there. And applying those skills creatively into areas where you have automation being adopted inside companies that's really interesting for me.

I see that that a lot of companies are struggling with things that to this audience might seem trivial, like processing invoices, it like automatically reading data from invoices, structuring that and feeding it to an ERP system. That's in huge demand right now. And if you have a good solution for that you can sell more like enough of it.

But that's like the trivial sounding problem. But I think I'm saying that, that the problem that you're going after, doesn't always need to be sort of fancy and high flying, sometimes more sort of simple thing can work. And also kind of thinking through that lens, when you're thinking about starting a company starting a business, it doesn't always need to be a completely white space without any competition, that's going to be a red flag for any investor out there that, you know, why isn't there any competition? Rather, you know, we have companies like UiPath hat has raised a billion dollars in funding, but it's a validation that this is actually a working market that we can attack. So, so that's the that's how we are thinking about it. It's good to have competition, and you don't need to always reinvent everything in order to start a startup company.

And I think the final piece is that you probably know a lot of things that like 99.9% of the world doesn't know. So don't discount the specific knowledge that you have in your specific domain.
Peter Zhegin:
Fantastic, thank you very much, Antti for this optimism, and for spending your time with us today. Really looking forward to hear a lot of interesting news from you and from Robocorp. So really, thank you once again.
Antti Karjalainen:
Thank you. It's nice talking.