Antti Karjalainen is the Founder & CEO of Robocorp, a San Francisco and Helsinki based company that set out to change the licensing and delivery model of robotics process automation with open source technologies.
We went in-depth on the state of open-source robotic automation with Antti Karjalainen aka Antti from Robocorp. We dove into Robocorp's upcoming offerings and the roots of Robot Framework.
Google Play / Apple Podcasts / Spotify
Google Play / Apple Podcasts / Spotify
Interview with Antti Karjalainen from Robocorp
SPEAKERS
Mark Percival, Antti Karjalainen, Brent Sanders
Mark Percival 00:03
Well, welcome to another. It's automic podcast this week we'rejoined by Antti from Robocorp. He is the CEO and founder of Robocorp. And we're going to get a little background on him and talk about open source RPA today.
Antti Karjalainen 00:16
Hey, great, thanks. Great to be here.
Mark Percival 00:18
Yeah, thanks for thanks for coming on the show. So I think you know,what, the easiest way to start off is just kind of get into a background on your background and RPA before we jump into sort of the larger issue of opensource RPA, and we would love to just kind of get an understanding of, you know, how you started in this space?
Antti Karjalainen 00:35
Yeah, well, you know, that's a, that's a good question to start with, it's gonna be a long answer. But I tried to keep it brief. also ties into why I'm into open source side of things. So, you know, my background is in software engineering, I've been doing it since like, 2007, I think I counted.And, and I got involved in a project called Robo framework. Early on in 2013.It was back then a test automation framework, and open source one, and actually had a company where we, together with some, some partners, focused on doing consulting services around test automation, CISCD and DevOps type of things.Early on in 2013, starting onwards, and, and the first time I actually saw RPA, being mentioned was at a at an event in 2016, late in the year. And, and the first thing I thought that when I saw that kind of concept of RPA was like,Well, isn't that exactly what we're doing in test automation? I call come is that any different and I kind of felt compelled to start finding out more, I started meeting with people who were in space. And, and through twists and turns, I actually ended up pushing robot framework, the open source project to become an open, open source RPA tool, eventually, that happened in 2018. But there's a lot of coincidences and a lot of opportunities that are lined.
Mark Percival 02:18
That's an interesting point. I do think, you know, when we think when you see a lot of the open source RPA tools, there is sort of a focus on testing. And you're right, if you come from a software background, you know,automated testing has existed and some format that looks like RPA, but has always been focused around more of the, you know, testing a front end testing a web, you know, implementation, that kind of thing, less around the automation side. So, you know, you mentioned robot framework, maybe give a little background on robot framework, because there's robot framework, right?And then there's the RPA framework.
Antti Karjalainen 02:51
Well, yeah, well, they'll do different things. So, so robot framework, the history starts way back, early 2000's, it was originally developed at Nokia, as a generic test automation framework written in Python, I think that was one of the kind of best parts about the origin of robot framework. It was open sourced in 2008. And then Nokia sponsored the projection wards, until 2015. And at that time, the funding was cut off by Nokia, so,you know, my company back then we decided as kind of partners of the founders of that consulting firm that, you know, we should, we should push this forward.And we should make sure that this tool gets developed further. So, we got together with some other people who are using the tool and, and started the robot framework foundation that supports the development of the project. That was in 2015. And, you know, I've been hosting robot framework conferences for three years now. We have Robocorp, every January, have had it for three times now. And yeah, so robot framework is kind of the core tool itself. And how it's designed is sort of generic automation framework, where you, the call itself doesn't do too much. But what you do is you integrate to your application that you're automating with, with something that we call test libraries. And one of those test libraries is now the RPA framework that that Robocorp develops. And that's an open source project is freely available on GitHub, if you want to go and check it out.
Mark Percival 04:37
Got it. That makes sense. And then looking at robot framework,there's sort of an idea this reminds me to like, want to say like cucumber,like it's a very as the language is supposed to be in some in a different format, or is it? Is it all written in Python?
Antti Karjalainen 04:51
Yeah, robot framework has a user base of, I know what the actual number is, that tool itself gets downloaded around 5 million times. Annual right now. This, this, uh, around 8000 people registered on the robot framework slack workspace that we have the robot framework or gets, the website gets around 45,000 visitors monthly. So, I'd say that the user base might be somewhere around hundreds of thousands of users. And I am pretty sure that majority of that is still on test automation, because the update use case is so new. We introduced it in 2018. June, we have the I think it was 311 release, or the tool that brought in the obvious index. And right now, we've been actually doing a lot of improvements on the core tool. So, we just got a new, new parser structure on the front end of the language, which allows for better language server protocol support this kind of technical thing by it just means better editor support in the end of the day, like VS code support that kind of things. So, um, so yeah, I mean, the other use case is definitely growing. And what's interesting is, it's actually bringing new people into the ecosystem instead of like, converting a lot of the old so I see that the robot framework ecosystem is gonna be growing actively in the coming years.
Mark Percival 06:20
Right. And looking at the RPA framework, this is getting more into the RPA side of things, right? Or it's like looking into things like Excel,and, you know, word and all the different desktop, you know, applications.
Antti Karjalainen 06:33
That Yeah, for sure. I mean, I mean, robot framework has been used to automate anything from mainframes, to dynamic web apps, to elevators to telephone switches, you know, so you basically, whatever you want to do with it, you have a library for that purpose. If you don't have one, you can write it up with Python robot framework has pretty clear and easy to use, integration API, you know, for building libraries. So that's one of the strengths of the tool is essentially, you know, anything that you have working with Python, you can integrate it with it. But it still provides a bit more top-level structure than using Python directly to automate things. But yeah, like you, like you said, you can do like SAP stuff, already Salesforce, you know, what have you this, there's a ton of different projects that the community has built over the years. I think, if you go to search for Python packages, I mentioned robot framework, you're going to find like three to 400 different results.
Mark Percival 07:34
Yeah, that makes sense in and looking at this, let's, I think we useful to back up a bit and talk a little bit more about sort of the opensource RPA versus what we have day, which is popular in RPA, which is things like UI path and automation anywhere, just these large platforms. And obviously, you know, I think we could talk about this a little bit later with Robocorp is doing but overall, there's a, there's a, you know, a large number of people that are using these platforms out there to do RPA. And open source still seems kind of new, it still seems like people are kind of defaulting automatically, if we talk to people in the RPA space, they kind of default automatically to one of the large vendors, and open source is just not yet on their radar. He talked a little bit about the difference between you know,where you see open source kind of coming in and capturing market share from these guys.
Antti Karjalainen 08:22
Oh Yeah, you know, let's even start with that. I think, I think fundamentally, there's nothing that will prevent it, prevent the RPA market from eventually being swept away by open source technologies. I really do believe that there's nothing too defensible about the proprietary tools that are out there, on the on the technical level. So, I see that actually, they're the individual software, but as a concept is going to be commoditized. And the pricing is driven to zero eventually. And, you know, there's still ways to go for open source RPA, for sure. And we are part of that process at Robocorp. We try to make it as easy as, as easy and as compelling as possible to start using open source technologies to build RPA solutions. But ultimately, why do we even want to use open source? Like what was the point of all that is because I think personally, that we need to have a kind of open ecosystem may not be something that's open for sharing and collaboration. So, this comes back to kind of my original realization inside the RPA technology scene was that, you know, I, I saw when I started going into understanding RPA in in, in 2017 when I was exploring that, actually. So that is really a fundamentally a developer domain in the end, when you talk about RPA the sense that you're automating back office tasks, Is this actually a developer space, not a business user space regarding the tools. So, if you treat it as a developer space, you kind of understand that you need to have some sort of open, common ground, like a set of technologies and tools that are open for sharing and collaboration, and then build the industry on top of that. So, I think, you know, that that drove initially my thinking around open source RPA.
Brent Sanders 10:28
It's interesting, because it's a, you know, we talk about this on nearly every podcast, but this, we believe, to be sort of a fallacy around the,in the community driven development from within an organization, which it's,you know, it's more of a theory. And we've talked to a couple people that have said, you know, you can support teams to do actual create automations from the business. However, I just don't see a world in which that that scales, and will, we've heard the most is around, you know, once there's a critical mass that has hit that or that arrives, that scaling up the automation department,you eventually start turning it into an engineering department. So, I would agree that it from what makes sense to me, and at least my background, and what I've seen is, it is fundamentally developer space. So, you are writing code,you are doing all these things that sure you're doing it for the business, but how is that really different than any other part of software engineering?
Antti Karjalainen 11:31
Exactly, exactly that, that. That's my thinking exactly. And, you know, I think it's not that different, that we talk about a Robocorp we talk about software robots as a concept, we like that term more than RPA. And we use software robot developers, as a term to describe the people who will do that build software robots as their main job and maintain them. So, its kind of better captures the nature of the work that these people do. And we want to enable the software robot developers with proper developer tools they will actually like to use in their work.
Brent Sanders 12:13
You know, you look at some of the tooling outside of the opensource space, that you're creating software robots, but you're doing it using tools that don't seem very conducive to complexity or conducive to rapid change and rapid development. I'm curious, you know, what you've seen outside of the open source space? And, you know, obviously, you you're coming from inside, as from a development or scripting perspective, I'm curious, you know, when you talk to people that are using robot framework, and more specifically, your product, do they? Are they coming from these sort of large third party vendors?Or are they starting fresh? Generally, from a development perspective? I'm curious what you're seeing, specifically around people that are getting started?
Antti Karjalainen 13:06
Yeah, well, you know, there are many, many types of people coming in some have experienced would be proprietary tools, some, some have experienced Python, some have experienced with the robot framework before. So,there's many, many points to start with. We try to give everyone tools where they somebody coming new. So, we have an IDE that's built on Jupiter labs, we call trouble code lab, which is a single click installable application that comes with everything like batteries included, then we have, we're going to be releasing a Visual Studio Code extension as well, that's for get to work more with the developer crowd. But ultimately, both of these tools they work with,with scripts, and text-based task definitions. So, I believe that, you know,text can be very expressive, and, and you can express really complex things with scripts, that that can be modular and build on top of other scripts. So that's a great way to do build an automation routine. I've seen some of these proprietary tools like visual editors being worked. And they do to me seem needlessly complex. If you come without, you know, if you can't be just programming experience, you just look at this is a whole another world that you would need to learn and understand and master. And it's just so kind of how to even approach that. It's, you know, it's, I think I have some sometimes-drawing parallels with the kind of the early 2000's, you know, attempts on creating UML charts that will generate C++ code. I just never will works that way, every,you know, every 10 years or so 20 years, we have this great idea that, you know, we should we should do like visual programming and, you know, in the end is a mess, you know, I get it that like low code, and no code tools can work and you know, you have this great tool such you can use to string together API's and, you know, do really powerful things with that, but, but ultimately,with software robots, how we see this is that you need to have the full expression power at your hands and have that capability to automate anything kind of the last mile of automation. You know, if you're doing like, just automating Salesforce, and some something else, like stringing those together,like, that's fine. I think there's better tools and RPA tools for to do that.But if you need to do your, your mainframe terminal based, you know, insurance application workflow thing, then then just, you know, use software for that.
Brent Sanders 16:01
Yeah, and the thing that you sort of get for free, it seems like you correct me if I'm wrong, but using a platform like robot framework, you have everything that's at the disposal of Python. So, when you go, one thing that I always am thinking about with, you know, software robots in production is how do we monitor them? How do we, you know, where our logs going? How are we using those logs? How are we managing exceptions? Well, you know, in this world, I think one of the key advantages, in my mind is you can tap into everything that's available for Python. So, using services like century or, you know,streaming your logs. And, you know, I'm sure you can do that in proprietary platforms. But it seems like it comes with a cost, right? It's usually, you know, having a certain license level or having some of the infrastructure pieces that are recommended by that platform. So, for example, like how do we how do we get our logs out of UI Path versus out of Python? It seems straightforward.
Antti Karjalainen 17:05
Yeah. And I actually robot framework, one, one thing they provides you with is a uniform logging format. So, every robot execution is a creates this currently is XML based, but we are actually driving it to be in the future JSON based log format that stream so is really accurate description of the robots activities, did you get like free from every execution run. And, and I think we haven't really touched on that on the aspect that we are working with,at robot, Robocop, but, you know, we want to, you know, really enable you to focus on that building pod, which is like, super good and nice with these Python based tools, but then leave the infrastructure part to us, which is the orchestration side of things. And that's usually where, where you need some additional help with working with Python based stuff, especially, because Python isn't really well known to be that, that kind of portable thing on its own. So, it's actually really difficult to deploy anywhere.
Brent Sanders 18:17
You know, with Python, there's generally a, you get a runtime setup, or it sounds like you guys are taking advantage of Jupiter notebooks is that that's correct assumption, as you basically have sort of this one click batteries included environment, you can get your sort of local build running your automations running in then I can take a Python script or two or However,their LinkedIn maybe a zip file or a collection of files and upload them to your platform. And they will run and I'm thinking this sounds to me something like a Heroku for RPA.
Antti Karjalainen 18:52
Well, yeah, kind of. So Robocloud is the orchestration platform that we're building. So that's essentially pretty much the same thing as you would expect from an enterprise grade proprietary solution. And how that work is that you build the robot with a with either Robo code lab, or then you know,Visual Studio code, we have command line tools to support workflows for people who, who don't necessarily want to use something like Jupiter. But for anyone new coming into this ecosystem is really easy to start working with our Jupiter lab-based robot code lab. And just install it and you will always have the latest and greatest up to date stuff available. But you know, when you're done building the automation routine, or you download somebody else's ready made thing, you can then just use one command line command or press a button on the lab, and then just push it to cloud. And on cloud, we have you know, kind of basic concepts of organizations underneath organizations we have workspaces So your company can have an organization that, you know, has multiple workspaces,you might have testing and production, you know, you might have one workspace for IT and one for HR for their bots. And then then you have basic secrets management, these kinds of things on the cloud as well. And then these, these spots that you build become activities on the cloud, and you can string together processes from the activities, and they can be run concurrently. You can run them through our worker process on prem, or on a cloud instance. Oh,yeah, it's pretty flexible platform interesting. And what that solves for you is, is all the, you know, messy things about pouring on, like having Python be portable. So, the worker process that will actually execute the process that handles all dependencies, all custom dependencies that you need in your process, and even if you write custom libraries, as well.
Brent Sanders 21:02
Interesting, so, you know, worrying about, as you mentioned, you know, it can run on an elevator can run on a mainframe is you will basically encapsulate everything into some form of runnable thing that can be downloaded to your local on prem environment and then executed, like if I am automating, you know, human resources, software, or payroll software that's on our in our data center. Yeah, it can be downloaded sort of to a local machine. And but you guys.
Antti Karjalainen 21:36
Well, it's really seamless the way works, you know, you do the automation,like you have access to HR software, right. So, you, you write the script, you test it out, then you press a button, and that script gets pushed to the cloud as an activity. And then, you know, when you want to run it, either locally or on your, on your laptop, or on a on a virtual machine somewhere on your private cloud, you will just install the worker, worker process over there works on Linux, back windows, basically anything, and you log in to your RoboCloud account, and you can start running the process. And it doesn't need any firewall openings or anything like that just one port open up, like HTTPS connection out outside. And you're done.
Brent Sanders 22:23
It's interesting. So, one question I had around, in general, a more of a tech space solution, right? So be it open source or otherwise, it's like,how do you feel like I should back up, you know, one of the things we talked about often is how brittle RPA is, you know, by nature, you're working against multiple systems, and things may change, and therefore, you know, bots can go down. And that's sort of, we spoke to somebody recently, and they were saying RPA is different than regular software engine is largely non deterministic.Now, there are times when, you know, you're working against systems, and they just may not be available, they may not be ready, and you have to sort of deal with that as best you can. So, I'm wondering, you know, how you see maybe text tools or open source tools being more or less responsive? And what effect Do you see that it has on responsiveness and being able to adjust to change?
Antti Karjalainen 23:19
Yeah, well, I think there's few things involved. The first thing is that we, we kind of focus on software robot developers, so the kinds of level access level of excellent expectation is such that we are working with developers who have basic understanding of, of things like error handling. So,and we actually teach people this stuff. So, we kind of teach the basic concepts how to deal with exceptions in the first place. Second, I mean, gee,you know, you know, what really changes a lot is, is a piece of software that is under development. So, the things that software, like Robot framework is built for test automation. That's, that's basically working with software testing software, that's, that's constantly changing. So, so you can see that,that that kind of tool is really suitable for RPA use, where you have this need to be able to respond to changes in underlying applications. So I think, I think, like, it's, it's part it's, it's about conventions, and part about the kind of the capabilities of the tools and, you know, I'm, I'm kind of on the,on the fence on the, on the idea that that should be should we try to do something like, you know, teach neural networks how to identify, identify buttons in in UIs, I think, you know, we basically have this idea about doing that as well as an open source thing. But, but then again, it's certainly not like deterministic anymore. If you if you using like locators and, and going through API's where you have those available and using best practices and building it robustly. It's, it's fairly solid, to be honest, yeah, there's not too much breakage, you know, if you if you go and do like image driven, like image-based recording, for sure, that's gonna break a lot.
Mark Percival 25:21
So, um, quick question around who makes this decision to actually,you know, use an open source RPA tool, versus using a platform, I think we talked to people in this space. And a lot of times these decisions come from somebody like the CFO, and so the CFO will basically come in and make a decision to look for some cost savings with RPA, and will then choose a platform and then choose a development company or a consulting firm to come in and do these implementations. And, you know, a lot of times some of this stuff is driven by the developers. And this case, it seems like it's more likely that the developers are going to be the ones that kind of want to push this as the as a solution versus, say, another platform. And it feels a bit like, you know,around 2009, I was working for a company. And I remember the developers are really very adamant about using something like EC2 or a cloud computing framework. And it was they were pushing it up to the management level versus the management pushing it down. Do you see something similar here happening? Or like, I just, I guess a larger question is who kind of makes the decision to use something like open source? RPA versus automation anywhere?
Antti Karjalainen 26:30
Ah, yeah, that's, that's really good question. And there's no really clear content answer to that one, I think I'm pretty sure that over time is going to be more developer driven. That's the thing that we, we really believe on. So, our focus is on developers purely. We don't, we don't that much, at least at this point of the company's life cycle. We don't talk too much to business users, or they're kind of certainly I'm not talking with the CFOs. spending all my time on that. Not to say that there would be any anything wrong with that, but I think it's, you know, their, their CFO who, who wants to do those cost cuttings. Firstly, obviously, our beer has much more to offer than cost cuttings, but if that's the driver, they probably turn to some professional services companies and the company and, you know, ask them to automate the stuff. Well, it's really up to the company to recommend a tool that does the best job and, and if there happens to be robot framework and Robocorp suite, then then they're probably recommend that kind of, you know,that would lead up to a discussion around our business model, which is, I think, pretty unique in the space. But, but still, like, who makes the decisions, I'm, I'm actually betting that, you know, when we, when we have saturation, around open source technology in the RPA space, and really, some of these technologies start maturing and becoming viable alternatives to these complete solutions that are out there as proprietary, we cannot see the same thing happening as happened to let's say, databases, while ago, you know,previously, when the when the CTO wanted to purchase a database solution, they would call in a salesperson somewhere, and the salesperson would fly in from a different country and take the CTL just take dinner, and, you know, spend a few months preparing a quote, and then they would, you know, sign a deal for 2 million dollars for, you know, database services, and tools. You know, right now, if you need a database, you just use MongoDB, or whatever is available developers to make the decision, and then they force the CTO to actually accept the purchase of MongoDB Atlas eventually. So, I see that RPA is going to go the same way. There's really no reason for the seat, CFO to be even thinking about what technology they need to adopt, they just want to have the results.
Brent Sanders 29:07
Yeah, I, I agree with the perspective in the sense that like, if you leap ahead, because we're already seeing license costs go down, we're already seeing more players and everything become more commoditized. So what you said earlier, you know, the beginning of the podcast, so sort of this race to the bottom, or race to zero it, it's going to happen just because if people are seeing dollar signs now and so they're saying, okay, we're gonna create our own studio, and we're gonna go, we're gonna do the same thing. UI Path does,but half the price and we'll throw in an extra feature and then somebody's going to keep it's still going to be worth it for these companies to create these tools and they're gonna figure out ways to other places to source or create revenue. And so I, I see the vision, you know, we'll see what ends up happening but I completely I can see where you draw these conclusions and looking at in the same sort of thing that we keep talking about, which is this is a space that feels like it's, you know, where the web was, you know, 15 years ago or so. And I do think that there's a lot of similarities and a lot of parallels that will happen will be interesting to see what actually happens I,but yeah, I totally agree where the more and more companies will start to adopt our RPA in general, right, this idea of Hey, we're going to have maybe it was it was piloted by the CFO or piloted by the finance department. But eventually,there's going to be an element of it, or an interdisciplinary group that has software engineers on the team that is managing the whether it's proprietary,and what happens in is they're eventually going to start moving it into tools that they're more comfortable with. Right. And so there, we spoke with a talented dotnet engineer who's managing UI Path right now. And so, it's like,he hasn't touched visual basic problem in quite a long time. But he's now in charge of five to 10 scripts, and it's, or should say, processor activities.And, you know, what is it going to do when they start breaking in a month, six months or a year from now, he's going to revert to the tools that he feels more comfortable with, and likely not going to be the Visual Basic visual editor,it's going to be something. So, if that's the vision, I'm seeing, I'm seeing a lot more automation practices that are popping up. And I think there are going to be automation departments in the next couple of years that are going to just choose the tools that are they're going to choose the Mongo versus the, you know, Oracle Database?
Antti Karjalainen 31:50
Yeah, yeah. I mean, I, I spoke with a with a talented RPA developer, I think he was from Spain, from Madrid, and spoke with him and, and he told that his job for the past two years has been to fight against all five fight against blue prism, I think the first thing they have saying, you know,he was wrangling the tool to be able to do the things that they needed to do.And again, I'm, I don't want to, you know, name names on the on the proprietary side of tools, I know that these tools have created this industry. And I think that, you know, they really shown you know, that there's just a lot of value here, the kind of the promise of RPA, you know, is so simple, like, if you can document something, you can automate it and, and I think software robots as this agent based automation solutions, they really have their place, and they can provide so much value, when you're in the right ways that, you know, or, or married the companies who have dreamed the industry forward, but I think it's coming to a moment where we, we gonna have different kinds of approaches be more prevalent, and different kinds of business models. So, I think they,they're thinking around pricing, RPA, robots, I think it's a, it's a huge mistake in the industry, I kind of get where it's coming from. So, you, you kind of do this calculation that this bot is going to do one person's full-time job. And so, you can, you know, spend 20% of that person's salary on the bot,ISIS, but it ultimately drives you to use less automation, not more. And I think we need to have some innovation, the business model side as well. So that's, that's why we don't price per byte on our side. At Robocorp.
Mark Percival 33:45
That's a really interesting point. I mean, I do think that that is exactly the issue is it's about utilization. So, people immediately start looking at, well, how many bytes are what are actually how many bytes we actually need to run these processes? And then how do I start to cut that hair that back?
Antti Karjalainen 34:02
Yeah, well, I mean, I mean, the first time I got into a sales meeting, I was working as a consultant and just teach my way into a sales meeting talking with our pay clients. And I think we're asking exactly these kind of questions like, how many bolster journey to buy, how do I make sure that these bots run like 90% of the time? can I build some cues for the boss so that they can, they can be like, highly utilized because I'm making this investment and I want them to be, you know, they were optimizing them that the licenses similarly to like production line robots, and didn't make any sense tome that the marginal cost of the software is zero. It shouldn't be like that.So. So that's when I realized that we need to have a different business model,in addition to different tech technology approaches as well.
Brent Sanders 34:53
Can you tell us a little bit more about your pricing model?
Antti Karjalainen 34:57
Yeah, for sure. So, so we have basically three products, which are,you know, Robohub is our learning platform is it's a site that you can go to and browse around and just learn about software robots in general, obviously free to use. Then we have Robocode, which has the command line tools and lab and Visual Studio code. And that's the best stuff that's free. And we're going to be open sourcing all of that, actually, after we go the GA release. And then we have RoboCloud versus the cloud orchestration platform. And on RoboCloud, we have a free tier for developers. And then we have a T tier for teams, and, you know, team tier for bigger organizations, etc. And as always start with zero,we go up in the subscribe subscription prices, and it's a flat monthly fee,basically. And you can you can use on the on the pro tiers, you can use as much as you want to use. And on the lower tiers, it's basically limited by the sizeof your team.
Brent Sanders 36:04
Right? Yeah, cuz I mean, I think there are even on the proprietary,and it's, it tends to be providing this utilization problem, which is just doesn't. It's confusing, why that's, that becomes a new focus of a team that's trying to get the most out of their licenses so expensive.
Antti Karjalainen 36:22
Yeah, that's, that's so harmful for the for the overall perception of RPA Think about it, you know, try to come up with a better, better pricing model, we've been thinking about this a lot, as you can imagine, but, you know,it's, it's really sort of tempting to go there, because you think that you know, this is the unit of value, the more the execution time, maybe you might do like time based, but then again, people start optimizing on, on the time it takes for the robots to run, probably unnecessarily, so we don't really want to go there either. So, you don't want to go anywhere, that will actually decrease the motivation of developers to interact with your tool. So, you want to have all the developers but for free out, out there, obviously. So. So we think that the orchestration platform, if we can provide it as a as a fully hosted service,the super easy, super compelling to use, and really saves your time, compared to trying to set up that kind of stuff locally, you know, you know, install an SQL Server and an Elastic Search and what have you, and, you know, forget about that, just start using it. And I think that kind of value that we can provide me that the cloud native platform for orchestration does, that's going to be super compelling. And, and people are going to want to pay for that.
Mark Percival 37:50
You know, we hear a lot about this kind of goes back to wait, maybe as rehashing, but we hear a lot about citizen developers, a lot about the idea of a developer, the kind of people come from different spaces in RPA. So, you find a lot of people come from the process side, and they jump into RPA,because its kind of fits there. That's their that's their job. Is this going in automate optimizing a process and automating it is included in that. And so, a lot of the platforms have been really selling on this idea of everybody developing a bot. But obviously, if you kind of go towards the Python route, it's looking like more like it's and it's sitting in the development shop, side of the business. What do you kind of do you see that this is the case this is going to still remain in the developer’s world? Or is this something that it's going to get simpler and easier, even from something like Robocorp’s framework for somebody to come in? Who doesn't have a strong development background to kind of jump into and start automating? Or is that a given a goal?
Antti Karjalainen 38:43
Yeah, well, the robot framework itself is has, you know, really low barrier to entry, it's quite simple to understand from the beginning. So,provides you a lot of kind of scaffolding, you don't need to be a Python developer to get into that. So, I think it's really approachable in that way,but it has a high ceiling so you can really grow as your skills progress. So, I think that's part of enabling more people to come into the industry. And to be honest, there's going to be a ton of jobs created for software robot developers, I think it's going to be a pretty big boom if we if you think that we're going to have tools out there that you can use for free to automate stuff or companies’ really powerful stuff. So, I think we're going to be seeing a huge increase in the in the amount of software robot developers out there and we need to be able to support the learning of those people. But in the end is you know, I've seen some, some tools advertise and this is not my wording, but I've seen this being advertised that some RPA tools they claim that they're suited for low skilled IT workers and I think that's the worst thing to say,you know, usually, yeah, yeah, yeah, you need to have skill in order to do that, for sure, you know, to be, like successful. And, and, and I, I don't want to go around saying that, you know, everyone should start building up as and as an analyst, for instance, I mean, there's different roles, and I'll be for sure. So. So, you know, I'm pretty mixed about the whole idea of a citizen developer, you can be a software robot developer, and go for it. Or you can be a proud analyst and go for that. But, you know, not everyone has to be a software robot developer.
Mark Percival 40:38
Yeah, I made it as it does sort of reinvent the wheel, right. I mean, we already have this idea of a product manager going in and laying, you know, documenting a process, and then software developer’s kind of coming in and tackling it.
Antti Karjalainen 40:49
Yeah, yeah. And I mean, that this goes back to talking about the visual programming aids, you know, that the next thing that gets reinvented every 20 years or so it's like, you know, we could build a tool that the manager can read this process description, or test, the script, like who manager in in history has ever wanted to read a test script or process description? Like, no, literally no one? So that kind of reinvents that as well.
Mark Percival 41:22
Yeah, this feels like I don't know, Brent. I feel like this is goes back to old cucumber, Ruby days.
Brent Sanders 41:30
Yeah, well, in the Python, where we call that lettuce, lettuce?Yeah. Sure. Remember, that was the writing of gherkin, right, right. It was a really great, approachable way that you get a project manager or product manager involved in testing. It's like, Hey, you know, better yet, how do we communicate requirements, you know, given when then, it was a mantra that I don't know why it was so popular, I think around 2014 or so it was very, very popular. In some practices, it seems to have waned. But if I ever need to just boil down what the thing should do, that's, that's my go to way of you know, or even filling out a ticket and saying, Hey, this is something's not functioning correctly, given why I did this. And given when then, is a powerful set of statements that you can use, and so I can see why you could.
Antti Karjalainen 42:21
Yeah, actually Robo frameworks bots has index by the way it has gherkin syntax. So that's supported in there, but I've never been a fan of that, that, you know, trying to make everything understandable. format and for managers, kind of So, you know, you need to have the expression capability as a developer, I've seen Robot frameworks index is actually plain text readable. So,you can you can write it, you know, the best role framework tasks that are out there, they have written as a story that you can read in clear sentences. And that's, that's really nice. But that's on the most top level of the syntax money, really into those key verse underneath? you'll, you'll see a lot more technical stuff, then.
Brent Sanders 43:08
Yeah, it's very interesting. How much of you know, it's, you know,coming from the, the testing world? And again, most of the use cases, five, six years ago, it sounds like we're around testing. I mean, how much of that is informed? What are the what are good candidates? for, like, what types of things can you automate? And what are maybe less so because you know, another thing that we see with most platforms is you want to tackle your low complexity, sort of we call low hanging fruit, these low complexity items that are taking up man hours, I mean, is that's still generally the case for, you know, open source RPA, or an open source was specifically with a bot robot framework. I mean, do you see that as being a good thing to go after? Or do you see differently?
Antti Karjalainen 44:00
Well, I mean, that's a category that way we can call that certainly happens and, and then then you, you have different types of categories. So, you have replacing, replacing that low hanging fruit routine work, obviously,something that you should do, I really like the cases too much where you can,where you just go in and replace like a bunch of routine clicking on your desktop. Because that's if it's tied to the desktop, then it's just making yourself a bit faster, essentially, like macro recording. That's, that's really not too powerful use of that technology, but certainly something that can be done. But I, I tend to always say to people that that tries to think about ways to actually get rid of the human loop all together. So, if you have a process where you need to go, you know, open stuff on your desktop manual, like could you just do it in there in the back office, like just have the robot deal with that and maybe send you an email after it's done. So, so yeah, I mean, some of the RPA tools out there can be used as an effective macro recorder for your desktop actions, we haven't been like gone out then and actually implemented anything like that in our tool set, because we don't really see that that's,that will be one of the most powerful use cases, but I might be wrong and a lot of users might start complaining and we'll do that's for sure, that's gonna happen. A lot of people want it. You know, just I think we, we as a company are kind of like being on the stealth side of things. So, so we haven't released publicly yet we are going public with our software this summer. So, it's going to be in July, when you can just go and robocode.com and download our tools and sign up for RoboCloud for free. So, so I think that's something to be on the lookout for. And the reason for that is that we've been, there's a lot of things that we've been wanting to build out properly before we go out. So, we want to have a great experience for developers and also have a lot of good learning material out there before we say that you can start using this and we want to be the premium product on the open source side. So, so nothing where you need to go into GitHub necessarily and start poking around and downloading source code. In order to get going when you need to have that kind of sort of level of finishing the products that that feels like a properly integrated product suite instead of something that we cobbled together just in a few months. So, we actually have 40 people building the product right now and we've been going at it for a while so I think it's going to be pretty exciting when we finally released them in in July. We have like a few hundred people using them right now. So, so certainly we're not building in the darkness but going into broad audiences in July is, is really exciting.
Brent Sanders 47:12
Appreciate the time Antti this has been great. I think this is a good first installment and talking to somebody on the open source side. It's this is yeah, really special for us.
Antti Karjalainen 47:22
Absolutely.
Mark Percival 47:23
Yeah. Thanks a lot.
Antti Karjalainen 47:24
Great.
Mark Percival 47:25
Thanks Antti!
Antti Karjalainen 47:26
Thank you. It's really fun.
Brent Sanders 47:28
Thanks for tuning into the its atomic podcast with Antti from RoboCorp, myself, Brent Sanders and my co-host Mark Percival. Feel free to catch the show notes on our website at itsautomic.com. Thanks so much for tuning in.
