On this episode of the podcast, Brent catches back up with Dmitry Karpov the cofounder and CIO of Electoneek.
Google Play / Apple Podcasts / Spotify
Brent Sanders, Dmitry Karpov
Brent Sanders 00:06
Today we have a return guest with us. Dimitri Karpov, co founder and CIO and Chief Innovation Officer Electroneek. So Dmitry, we had a conversation, I have to recall it probably two or three months ago now. Since then it looks like Electroneek is rising in the market and growing things seem to be going really well. Maybe give us a quick update on your end.
Dmitry Karpov 00:33
Brent Sanders 07:10
Yeah, I have to pause you there. I mean, because it sounds, this is a trend we've been seeing happen slowly, a little bit, you know, with the open source, automation market, specifically, the folks at rebel Corp, we saw, they were leveraging sort of the robot framework and Python ecosystem. And the way I think of their company, and they may not agree with it, but like the Heroku of RPA, right is give me a platform where I can deploy my bots and have them run. But you know, moving backwards from there or at tries to say fast forwarding in time, from, you know, their inception and seeing where as you were talking about, like the UI paths of the world, or some of the larger incumbents, where there's huge license costs, there are the need for technical individuals to own projects and sort of be accountable to SLA's. So just to jump ahead and recap, though I couldn't agree more that this has been the trend. And this is sounds like the realization you're seeing it in, in reality is in terms of sort of product market fit of people being interested in, you know, lowering the cost of entry, and just having the right tools and being able to easily deploy the bots because I think, and I guess you tell me, I mean, I have not gone in depth, and I haven't used electronic myself, but it sounds like you know, with having an ID, you can write your own code, build your own automations. And this may not necessarily be code necessarily, as much as you know, working from an existing toolset. But the challenge becomes how do I then deploy those in a stable and common environment on my client? So putting myself as an agency? How do I get these on my clients, infrastructure or system? Or, you know, they want everything on prem? How do I deliver that? So I couldn't agree more with the direction. And it's really interesting. So walk me through, you know, when it comes to moving away from the enterprise, as you said, these projects get frozen up, but walk me through how you really pitch to these in find these sort of, not really the mid market companies, but you're seeing more of these entrepreneurial or agencies or smaller consulting firms. What has been the strategy for connecting with them and educating them on your product?
Dmitry Karpov 09:33
Great question. So I'm in charge for marketing, Electroneek smile on the back of my head. And I think marketing of RPA in general is very challenging. Because for a bigger PA for enterprise RPA, you need to beach, one of two things, cost savings, cost reduction, or innovation. What's the problem of doing both or any of them innovation, it's a hype so you can teach innovation once in the beginning. Get into the enterprise, make PVC, maybe build 1500 bucks, you wouldn't, you wouldn't get to large scale digital transformation powered by RPA. If you solely sell innovation as an idea declined, so you just sell cost savings. And that's where the rubber hits the road. And building business case for every board was pretty much what the reality of this industry, we flip the coin, so we don't sell to end user cost savings. This pretty much exists in the air, they realize the want to cut cost the therapy. But instead, we sell the tool to make money to 80 integrators. So we change the marketing language here to say, come and make money from our pa business serving clients and SMB mid market. Because now you're acquiring the tool to make money, not to save money for your own company, we don't need to do business case for disintegrators. To do the business with, they pretty much usually companies that work with us, they did CRM integration in the past, or something similar, maybe some accounting system. We, for example, have big success. Now working with Sage partners, we have this partnership with sage, second largest vendor of accounting software in the world, so many angles to that. But it's specifically interesting to see that there are partners who implemented such solutions to clients, they see so much demand for automation, that RP organically becomes another source of revenue for them. And we want to work with companies like that. So we want to give the tools to the hands of entrepreneurs. And entrepreneurs, when they think about tools, they, of course, want free tools. But at the same time, they understand that they can differentiate and pretty much charge a premium for the service. If this tool will be not accessible to everyone who just wants to get access on the street. It's pretty interesting that historically different RP vendors who became let's say big four vendors right now that have different politics, Silver's partnerships, they pretty much it was moment when you or Beth was given free licenses, prepartum licenses for everyone to everyone, then these times are gone. Currently, all these vendors they tried to kind of work with smaller numbers of bigger partners can provide the best price for enterprise. And typically their models will discriminate smaller partners, because they have a revenue share agreement. The discount you get for reselling, and RP license to end user will be bigger dependent on the volume of how much you sell. And in this case, if there is a competing, the two three competing firms, the bigger one will always be able to provide the best end user price for RPA technology. And because of that smaller vendors don't even give a try. Often many projects are on hold didn't fly. And at the same time, it's very important for those who build bots to build many bots in RPA business and I'm big proponent of that 80% of value comes from a developer not from the tool itself. But you brought up earlier how bots will be deployed in client infrastructure. That's an experience of RPA implementation team. That's their big value add, er B integrators don't make their money on development, only development is part of their business. But they also will help you to think through all the questions related to successful sustainable, stable, scalable, and scaling implementation. So, that's pretty much it. Marketing is very different. It's a very different conversations. And what's very, very interesting for us, this company is they build a lot of bots, because they have expertise, they are offering the game for many, many, many, many small projects. So be sure that our clients have a good service, because they work with companies that build bots at scale. And they build lots of scale because they don't need to pay their bot. So it's kind of product compliment and business model business model complementing product and marketing rep setup from negative way of thinking, reducing costs to positive way of thinking, making money. Yeah.
Brent Sanders 14:05
Yeah, I've seen a lot of that feedback, even in our practice around consulting is approaching automation with this, you know, we even have a I think we have an assessment tool that we built at some point, which is, it's kind of laughable now, this idea that, you know, give me the number of employees, their salary and the time they spend on a task, and I'll tell you how much I can save you. It's that that pitch really is not accurate. Andwe get that feedback constantly. And we obviously don't use it anymore. But the switching that to not what will save you But what will make you is an interesting shift of the paradigm. And I think that's more so the reality. I think what we see in successful automation is we are enabling the business to do something it just couldn't fathom doing with a person and that's usually more so the case. It's not that we had somebody actually manual Going through 10,000 emails or 10,000 rows or whatever, you know, massive amount of data manually, these wouldn't even take it on, you know, at times, you know, maybe you could offshore some of these tasks and imagine you're going to save this, but even then you're going to have a smaller sort of dollar per hour savings. It's just, it's not really the right way of thinking about it. So as you look at the platform, moving forward to the, you know, moving into the SMB market, which I couldn't again, I couldn't agree more with that sentiment of where automation is heading where, you know, kind of getting into all the different nooks and crannies of industry. Are there any segments that you're particularly excited about? I mean, and when I say segments, I think like RPA, and automation starting largely in finance departments. And growing up from there. Are there any new applications or areas of automation that you're seeing that are interesting?
Dmitry Karpov 15:53
Thank you for the question. So SAS becomes very important for us, because we work with smaller clients are more open to using SAS products. Let's start with finance, accounting systems, most of them will have cloud based accounting systems, SAS products. So it seems that to be great tool for our target market, we also need to be very connected. And we made a decision to move a connectivity layer from bots to orchestrator. So our Saturdays traders are the most think efficient on $1 will value in the market can be integrated with Zapier is big project for us. No zippers are the Y Combinator, the company recently got very interested in use on their relation. And they have 2000 apps inside them all connected. So to move anything inside zipper ecosystem is fairly easy. The bigger question was how you connect this system to RPA. So for example, something happened in your accounting system, you want to trigger a certain action in your, let's say, project management system, that is your legacy software. So you cannot easily connect it through API, or maybe API connectivity is very rigid here. So we integrated these two things. And now we are enjoying seeing very interesting use cases. For example, when you think when you think about DocuSign, it's something that is about dogs in the cloud. But at the same time to move to DocuSign. There is so much work happening in your Word files in a lot of work and DocuSign triggers a lot of processes inside your word slash Microsoft Excel ecosystem. And we wouldn't even like we would never be able to come up with such cases unless we built this integration to start to see what users do with that. And for us, it's another way to pretty much provide more value and make more money from the segments. Because we see now it's very important for us once we did the integration with Zapier, we see even smaller companies going with orchestrator from day one, or going straight through from months one to two. So that's something that we found very interesting insight, for example, with other vendors. Typically, if you don't buy orchestrator in the beginning, you will do it next year when you knew the contract. And you think okay, I will build more bots this year, I want to have admin tool or control tumor or something like that. Now our clients are about dozens of bots, not like hundreds. And currently, we see orchestrator being a part of standard package and 80% of deals, because it complements all the RPA capabilities with sales automation capabilities. So we think this technology focus is very important. And orchestrating for us is very important because of the infrastructure I described earlier, free bots mean that you will build more smaller bots and connect them and bigger workflows. And because of that, our clients will build this doesn't sponsor that will do a simple action for one to two, the easier to build. So our overall approach here is that let's let's lower the barrier for those who don't have great code structure. And when you do coding, typically you read the code more than you write the code. And reading the code is a fix is effective when you have great code structure. And I think we touched that on the previous code. There isn't so many RPA stuff on GitHub or get lab because the structure, if that's you want to change the status quo. Because we want people to write smaller bots and connect them together. We launched library of bots on our website a few days ago. It's not exactly a marketplace, but it's more a collection of very distinct use cases. Most of them powered by combining RPA with SAS. So go and check it out if you're interested. We plan to really source many many million watts from apartments, by the way, on the marketing side, so when you sell to the end user, and you want to source the success story from them, it typically requires you to get great excitement about the product, and some kind of expansion plans. So these are the most vocal users who are excited and want to expand their easy to ask to write a success story or use case, with our integrators, our clients, they are very motivated to publish success stories, because it helps their business and against make our marketing a bit more like a barrel rolling down the hill, rather than me rolling up the hill. So just want to draw attention to that. And the results.
Brent Sanders 20:42
Sure, sure, yeah, of course. I mean, there have to be the outcomes. Otherwise, you know, you're, you're gonna have a short time span. But it sounds like you're switching in and out. It's not really switching right? Or you're not not pivoting per se. But it seems like when we first spoke Electroneek, it's a platform, you're selling pickaxes. But I feel like you're making smaller, moving down market, I think, is a wise choice. I mean, obviously, it's playing out, and you're getting recognition for that. And it's great to hear. But the way I see the future of this space, or at least in the very short term is, and it sounds like you're in agreement is there's this new industry or consultancy that, you know, I think of and I come from sort of the software consultant background of you know, very early part of my career building websites, mobile apps, and then getting more so into web based software. And there were all these platforms that sprang up as my career grew, and the maturation of the technologies began to unfold, to make the job easier and easier. And so it seems like, to my point earlier, you're making pickaxes for people to build their own automations and just helping them build essentially a better pickaxe. And so as you look at, you know, how do you help people understand this new type of consulting and I shouldn't say it's new, right? It's, it's been around for a long time, but it's going to grow. And you're probably going to start selling to a less technical audience as you go. And that may be the way you were thinking and looking at, in contrast from electronic, which is, seems to be, would you call it a Is it a lower code or so sort of no code type platform? Or do you need to code in order to use it?
Dmitry Karpov 22:39
So that's a great question on balancing marketing messages with product roadmap. I don't want to say we are not a local platform, or that we don't do investments in becoming more power users slash citizen Automator accessible. That's, that's part of our vision. But when we see how the product is used by the consulting, let's say those who really have expertise that have to provide great client service, we start to think a bit differently about what our users want. And it's important for information, our roadmap, we see that ability to leverage custom code is very important for our audience, because that's the important part of creativity of developers. Right? So they want to differentiate on that level of creativity. And that's why it's important for us to make sure that whatever that consulting partner, or ag integrator want to customize for the best client experience, like connecting AI models, document recognition or something, integrating their systems using custom API's, or leveraging any coding language is important. That's more important for us than to simplify the functions we have. At the same time, we realized that the great way to build this is not the major is to limit functionality on this tool. So it's if we assume that the matrix can automate 30% of cases, then we need to limit the availability of functions in the product to the 30% of functions, and to do them really, really well. Because that's exactly what you would expect when you would say drive a car. If you are on a test drive for a car. You come in and you see there like a cockpit of Boeing 767. You will, you will be shocked. You want to have just one wheel, a few buttons and few dashboards to look at. So I became a proponent of it that one product, great product, cannot combine both. It can be serving simultaneously with the same product to audiences. So having different products for the segment's, I think is important. And allows flexibility for the vendor to actually make the product products the way that people want to use them. Sure, yeah. So, we still have power users. But we noticed analyzing Who are these power users in retrospective, most of them are data scientists, or they have some SQL knowledge. So typically people, many people who work with data, they know SQL. And when they realize that they need to automate SQL or pretty much to something else, they say, Okay, I will be using very basic GUI level automation for that task for SQL one, I'm knowledgeable enough. So for them, it comes handy. But for most of our users, they actually want to have more and more what they will call custom code features.
Brent Sanders 25:50
Sure, and diving deeper into kind of the mechanics of Electroneek. When it comes to things around the orchestrator, like you were talking about earlier, you know, building smaller, more, I would say, more precise bots, let's say, you know, it's a process that you're looking to automate, it has three to five human steps, and then maybe you actually break that into eight to 10, RPA, or sort of bots that are going to perform smaller, more discrete functions. And they're going to work together, you know, the challenges that we run into. And we, in our practice, we're fairly agnostic, we use UiPath. Use Robocorp, we use automation anywhere in the past, we use just pure Python, sometimes if it's a fit, I mean, it's really been all over the place. But one of the challenges that we see across all of the platforms that we've worked with is a uniform way to sort of dispatch to orchestrate, right? It's inserted for our listeners, like, how do we save some information somewhere, pass that to something else? Do it in a way that sort of stateless, that's not going to be a problem to rerun many times? How do you guys look at those problems of sort of data storage, managing state and, you know, keeping things from creating, like, duplication of data, you know, walk me through the orchestrator and why it's so great.
Dmitry Karpov 27:19
Let me start by saying that our first orchestrator was just scheduling to, because that was clients were asking for. I wasn't scheduled by bots. So we need a scheduler. And orchestrator is pretty much what's what is built on that scheduler originally. So the functionality of that with regard to export in changes in your IT infrastructure, like having duplicated orchestrator, for example, on prem duplicating orchestrators, that's not something we are strong at. So we have relatively narrow set of features, but is good enough. And actually, for smaller clients, it's excellent enough for our target users. So it's, we've been in this many enterprise conversations last year about orchestrator. And we see that thanks to our great competition, the bar is very high for what the enterprise orchestrator product should be with regard to being able to scale virtual infrastructure, being able to provide backups, but scheduling security logs, connectivity of bots, something we focus on do really well. So that's pretty much comes from the limitations of where we develop the orchestrator.
Brent Sanders 28:37
Sure, sure. Interesting. So, um, as somebody who's, you know, sort of a, what I'm understanding is these agencies or I say, agencies, but these consultancies are becoming your new target market, right. And their clients are ultimately who you were serving in looking at, you know, what their concerns are, and how they worry about their, you know, implementations and what their sort of problems are, how did that shift in coming from the enterprise? Like, obviously, we talked about, they want to add value, they want to be able to sell this through and obviously make money. You know, that's part of the commercial aspect to it. But the things that they're worried about, and where I'm going with this is more around like, they're usually the people that are accountable to companies, right? And there's sometimes an SLA involved, like, how does electronic help kind of keep them in business without having to add too many bodies?
Dmitry Karpov 29:47
So I'm still thinking about what exactly unites all our customers slash partners, this agency slash consultancies.
Brent Sanders 29:57
Some, probably some wide array
Dmitry Karpov 30:00
Brent Sanders 36:26
That's interesting. That's great to hear. I mean, it's, it's fun to as we've had this podcast and has been in this industry for a while, it's fun to see that success is arriving to companies that share this vision that the value is in this SMB. And there are going to be more and more of these sort of partners or new automation companies popping up, which in my mind it again, I keep going back to my history of my career, which again, was this explosion of, you know, web mobile, and, you know, just in general web based software development that that started to pop up as more companies saw the value and needed it and became more commonplace. But, you know, as you look at the challenges that these companies run into with their clients, as you mentioned, you have support, you're growing support, you're growing. Tech docs and training, which by the way, is kind of a Herculean task, I never realized, until I've observed some of the players in this space, put together their knowledge base and see some of the knowledge bases that are crumbling. And I shouldn't say knowledge bases, but some of these like platforms that could put up like the UiPath marketplace is one that you'll see just tons of broken links. And it's like, it's really, really, really resource intensive to keep to put up the information and then to keep it up to date for year after year after year as new initiatives come up. So I take a hat off to that. And so this is kind of a long winded point of getting at how you feel about some of the changes coming to the landscape of RPA and automation in the future. So like, one of the big behemoths in the space, obviously, is Microsoft with power automate. And I'm curious, like, how do you see that tool? Obviously, dynamics is really big as you talk about things like SAP and ERP systems and seeing dynamics having a huge market share. And people just being in general, a Microsoft shop and having so much, you know, already had been bought into the Microsoft suite for 10s of years. And they have such a steady foot in the space. Like, how does that strike you? I mean, was you looking at this year ahead in 2021? What do you think about power automated? How do you think it's going to affect electronics market?
Dmitry Karpov 38:50
We lost power or domain. And I think it demonstrates the need for connectivity between SAS and RPA. And it's great that Microsoft moved in that direction. And for us, it's important to know that as you said before, companies that are deep in Microsoft will likely will go with power automate, just for the reason because it's Microsoft product. But there are many companies that are not in Microsoft ecosystem and for many reasons, they don't want to be there. And they would not use power to me for the same fact because Microsoft too so we think that will be ecosystems of RBS has connectivity growing around the world. And likely Microsoft currently has the largest one. So if we combine Electroneek ecosystem with Xavier ecosystem, we probably the second largest one or maybe even more, so how many connectors Microsoft has in their ecosystem. Think around 200 something as I remember and Zapier is 2000 apps. And we know that Microsoft will be growing their connectivity and we will grow our connectivity, but we will do it with the help of Xavier with our business partner in this case. So it will be consistent play will be very interesting to see, over time whether there will be other ecosystems growing around here, maybe now SCP can be leading this favorite own ecosystem of products, and so on. But I think what is very important for the market, and we just see the first signs of that as we speak in recent days, but we see much more of that towards the end of the year. And in 2020. That's a conversion from one vendor to another one is easy as far as hiring a specialized consultant, or as we progress towards the future click of a button in specialized software. So a few days ago, I've seen the press release by a company called thing blueprint. They provide migration services from one platform to another one, they say, you use you assign, one to use, aka, give us a call us able to use your iPad, give us a call, and so on. And we did some research to found currently there is one technology to the desert, from subatomic products to power our domain. But it's easy because first of all, one of the products doesn't develop anymore. So it's kind of static. Plus, Microsoft controls the version of power automate. So if someone is not only owning the tool, but someone external is building, that connector is a software sorry, that migration tool, the software, they need to take care of all version controls, it's a bit harder. But we see that consultants now build some MVPs on the tools to complement their own services for migration. And this will be I think, the most important for the industry, historically, we had high switching costs. In the year or two, we will have low switching costs. And that will be amazing Irish major vendors to go through all their ideal procedures and any other big things they do before the trend truly emerges. Because that will be I think, a disaster for someone in the market. And we think that our model is freebuds completely aligned with that. And we tried now we establish these relationships with this new consultant firms that start to offer migration services, because you want them to migrate to electronic. Because they still need a development ID tool to maintain your new connectors. It doesn't really matter for us whether it's partner who does their end user, but we want to work with partners who do migration. And we likely will see migration taken care of by software products in the future. If I would be implementation partner with resources, I will now invested in building kind of own practice for Migration tools.
Brent Sanders 42:55
Yeah, as you look forward to, again, 2021 in the next couple of years, there any trends, you know, in one in particular that were kind of stricken with or mastering one trend in particular that, and I wouldn't even call it a trend, one aspect of automation that seems to be getting better than I was very skeptical of, a couple years ago was around sort of the AI and machine learning aspects and how you can apply it to automation. And I'm curious, are you seeing anything from your perspective? Whether it's use cases that are, you know, being used with your platform, or other applications of AI and machine learning that have been interesting or, or notable?
Dmitry Karpov 43:39
Of course, we see it's, it's, I wouldn't say it's surprised, but, um, awareness about these tools about, let's say, image recognition applications, about document processing, about speech recognition, speech to text, text to speech is very high in emerging markets, even among small companies. But what we hear from the field is that the word doesn't translate into a demand. So our partners hear from clients. Could you tell us more about how we can leverage AI here, but typically once they hear the answer, they don't buy. At the same time. I think important trend here will not be an RPA tool being says there will be more SAS services, they do some kind of intelligent things for smaller companies, for example, says tools to process documents, hmm, people become more and more comfortable with that. So when we looked at landscape, and we thought that maybe for a mid market, we need to own on prem, OCR own on prem document processing. We found that many companies in the market who have 1000s of employees are now comfortable with using these tools in the cloud says services and so on. That's very important for us. I think that at least How we feel lowers the pressure on our PA vendor to provide that infrastructure for AI, but then increases the demand for connectivity to these AI tools. And that's where we will be thriving. We already have many partnerships already activated, or in the pipeline, some small companies that provide new solutions for SMBs, some of bigger vendors from accounting to speech recognition. And we think that just helping their tools, finding the right use case, and ultimately feed the data to the tool, and from the tool is what clients expect from an RP provider.
Brent Sanders 45:42
Yeah, I think that's what I've experienced my own practices, I am not going to be on a per project basis, I'm not going to develop a model or develop infrastructure to support AI and machine learning capabilities. As much as I can use Azure, AWS, I mean, things like Abby, these platforms that, you know, have composite of results that I, you know, are going to be so much leveled up beyond what I'll be able to do on a per project. Even if I were to specialize, let's say I went into logistics, invoice processing, and I just focused on that, you know, maybe I could develop a model, but even still, the infrastructure I would build on would be likely on AWS or Azure, or one of these platforms that would give me the tools to, to run for and looking at, as you said, This newer sort of client base, that you're looking at your customer base that are probably going to be dealing with 10 1520 different clients in a given month, it's not really going to be the center of you know what they're going to be best at.
Dmitry Karpov 46:51
Or maybe they have a competence in that. And in this case, it's easy for them to serve 20 clients, for example, right? In Latin America, there are many digital docs that are really hard to do any kind of standard trained machine learning models because of the languages or quality of formats of the documents. And there are many sales companies that try to develop that solutions. Some of them are not exactly SAS companies that are consulting companies that say we develop this great engine for very specific text documents in Brazil or in Colombia, to be automatically recognized and extracted. For them, that will be absolutely okay to serve 20 clients providing combined offering, but for someone who doesn't specialize in that.
Brent Sanders 47:43
Excellent. Dmitry, is there anything else before we wrap up that you wanted to? To mention? Are you guys hiring right now?
Dmitry Karpov 47:51
We do, we actually did a lot. So please check out our Engine list or indeed pages, we have many open roles and more to come. As I mentioned, before we standing up flap regional teams, that's not a few people. So it's definitely a rapid growth year for us. Five bucks last year, we don't don't plan to no decrease the rate of growth. That means our company is about to double sometime soon. And because of that, definitely we have global roles, if you manage if you ran global communities, knowledge basis, support, Eastern enterprise. Now it's not enterprise, much wider market. We really want to work with those who share the vision that the market is big, and was just waiting for the right tools to come here.
Brent Sanders 48:45
Excellent. Dmitry, thank you so much for coming on and talking to us giving an update. Congratulations on the success of last year. This has been awesome to be an observer of. So, thanks again.
Dmitry Karpov 48:56
Talk to you later in 2021.
Brent Sanders 49:02
Yes. Yeah, we will have to check in next quarter here. What's going on with Electroneek.
Dmitry Karpov 49:06
Sounds good. All right. Thank you, Brent.