On this episode of the podcast, we speak to Doug Shannon. Doug is a veteran IT product lead that has been deep in the world of RPA.
Interview with Doug Shannon: RPA Practitioner and IT Veteran
rpa, automate, people, automation, build, bots, run, understand, roi, whatnot, self healing, fail, governance, talk, process, groups, interesting, agile, companies, intelligent automation
Mark Percival, Brent Sanders, Doug Shannon
Brent Sanders 00:03
Thanks for joining us. Today we have Doug Shannon with us. So Doug is an IT product lead. And you know, RPA practitioner, and part of the podcast is, you know, we have different roles all over the RPA spectrum. And Doug is somebody with a lot of experience in this space. And we're hoping to kind of talk about his practice, what he's up to, his history, where he sees the future. So, with that being said, Doug, thanks for coming on. And maybe you could just tell us a little bit about your background.
Doug Shannon 00:37
Yeah, thanks, Brent. So background wise, you know, I come from, you know, 20 years of IT. I can also go into other things about like, why I think that's important. But, you know, it came in from, you know, back early on from like, doing interns and most of us that IT did, and then moving into, like the Department of Defense, from there moving to like, companies like the large companies like Chevron, like out in San Ramon area of California, and then moving to like, Cisco. And then now like other other different companies, and whatnot. And so, although we're doing some really neat things, in RPA space, we've actually been doing that for about a good little bit over a year. However, we're doing some amazing things that are talking to other RPA practitioners. We're all kind of like, Hey, what are what are what are what is that doing? What are we doing here? So that's kind of like the gist of that.
Brent Sanders 01:32
Excellent. And what got you into the IT space in the very beginning? I mean, did you go to school for it? Or was it something that you were always interested in?
Doug Shannon 01:42
Yeah, I was kind of raised in it. Right. So my father was very IT oriented. And he was kind of raised with the old Commodore 64. And having our own network setup, you know, you know, you know, type, you know, using IPX and whatnot, you know, the original original stuff, you know, playing Command and Conquer, and, you know, teenager, so, so getting into networking was a was a key thing, understanding how the internet and Yahoo, you know, messenger, and all the chat rooms and stuff originally, and all those Yahoo groups that were out there, you know, their entire introduction to the internet was was very interesting. And so that kind of drove that continual progress of learning and saying, like, Hey, what's this do stuff? What's going on? And so I think, I think there's 70 going, right. So a lot of IT folks out there, we're all kind of like a jack of all trades, we all kind of get around, do different things, because we do whatever job needs us to do. And so in that, we we get to see new sides of infrastructure, new science, technology, and really kind of find ways to marry the two because, like, we're like we're talking about like, RPA, for example, ai automation has been around for, for the beginning of it know, how do we make things easier for the business? How do we bring that together? And, and our patients like kind of the, not the newer way, but it's definitely one of the most interesting ways and very central low code, one that we could talk about, as well.
Brent Sanders 03:07
So yeah, so you know, looking at your career, in it, it's been super consistent in the space. And so as you mentioned, RPA has been around for a long time. And in fact, we've, we've sort of gotten the lore of the space from a variety of people. And it just seems like you can call RPA or automation, which is probably a better term to use it, you know, automation is with the advent of the computer, I mean, right? It's like, we're automating math problems we're automating, and it's all different degrees in applications. So it's hard to put your finger on, like when this all started, except, you know, I do feel like there was automation anywhere. I had a very early version. And there were, you know, gonna say power automate, but that's not right. It's one of the windows tools that was out a long time ago. Anyways, the question for you, the question I had for you is, you know, what do you think of this resurgence? And like, why do you think it's come about that there's, you know, this new label of RPA. And this growing, expanding? market, and we were just talking before starting the interview with, you know, companies basically going public, right, it's a UiPath just filed for their IPO, which everyone kind of knew was gonna happen. It's like, what is this resurgence all about? Why do you think it's happening?
Doug Shannon 04:28
Yeah, I think some people may even say automation, and the Microsoft side started around when Clippy was coming out, or, hey, I'm your assistant, I can do these things for you. And it's always interesting. It was fun back then, and you know, Windows 95 and whatnot. So there's that idea of beginning automation is part of it. It's a central link to kind of where we come from. The Resurgence now is RP is as of as the beginning of this COVID you know, situation. Getting this year, the year 2020. It was, it's now almost become like a household name a lot of people that had no understanding about it and narrow, they say, Hey, what's this RPA thing? Or how do I do that thing? Or is that going to take my job? That question, and we as practitioners have to answer that. So it's almost a matter of, you know, it takes years for something to actually get into the norm, or the normalcy of like the media or just in people's psyches. Right. So something like agile and agile management and scrums, all those things have been around for a long time now. But more people are like, Hey, what's that agile thing? You know? So those questions have been right, especially thrown around, which is fantastic. But it's a matter of like, how do we teach these people now? How do we teach everyone so that everyone has a better idea? And so, and then that's kind of your time, like, maybe we just call it automation, right? Because, because in the growth of RPA, you've been there, there's intellectual RPA, there's, you know, there's all these dangers of hyper automation, and all these like things we can talk about and whatnot. But the more jargon and the more information we put out, and the more acronyms we throw at it, the more people get mixed up and said, Well, I thought you were talking about RPA. So how that started. Going back to that is because I really think, since RP has an audit, an audit mechanic in it, to where we can see those connections and really say like, okay, yeah, it was automating these things. But now we can really track Well, what were they automating? How are we automating? What are those logs that and really like, what are these bots? And what are these, these automated processes to have access to? And so now that we can define that, by using RPA, I think that's really where it comes in. So you know, it's just that, again, like the resurgence of understanding on it comes into more businesses now saying, hey, how do we do what we do? And what are those things go, and then an RPA stance, we go, Oh, hey, we did these things. This season? Here's exactly where they are. Go get them kind of deal? Yeah.
Brent Sanders 06:57
Yeah, it's really interesting to see, you know, the the embrace, I guess, I mean, I've experienced this more recently, by sharing more, like screen shares, so to speak of, to people that have never used automation, or maybe they've had some sort of automation, but it was just like, not visual. So it's, I don't know, I feel like RPA Currently, the speed factor of being able to like, see, some people are just visual learners, right? It's like you show them something, and they get it. And that that's been a recent thing for me is just to see, hey, you know, automation, in a sense is happening with everything, you use a web application, use a mobile application, there's stuff going on that Yeah, you probably could break it into emails and have a person to all this stuff. But you know, it's bridging, bridging the gap. And so with that being said, I'm curious, like, what were the first RPA projects that you worked on? Like, what did they resemble? Was it like bridging old systems or, you know, working with old human processes to make them robotic?
Doug Shannon 08:01
I think I think everybody's first we're probably and we could agree that the first we're always the ones that failed, because it didn't tell coming into it how to do this stuff, right? So a lot of the times it's, it's somebody comes to you with a problem, hey, okay, you're doing this RPA stuff, let's go automate something, you have the proof of concept, you have the the general like, hey, there's this catch five kind of things we want to try. And then you and then you attempt to do those. Right. So I think where it goes from from there, but I mean, so some of the first ones, let's see, we, you know, we kind of dived into some audit stuff, we've dived into some, some interesting, like business case items of, you know, the average user saying, Hey, I take hour to three hours to run this thing. And then I go send it out to people, how do I do that better? How do I save that time back? And so I said, Okay, cool. Let's automate that function. How do you do that? Let's build like the PVD. Let's see, step by step. What are you doing? How are you doing? And then who do they need to go to? Luckily, you know, using different automation tools, There's different ways to handle that. And so, for example, for UiPath, for example, and we talked about them, they have like assets, and they have orchestration. And so some of those are the key meaningful items out of their automation tool is because you can do that and you can make more resiliency and more fault tolerance type of situations where you can build in many, many assets to say, hey, go to these different emails, or you can add them on the fly even in production. If you had to say, Oh, we forgot to add this person to this email system.
Brent Sanders 09:35
Yeah. You know, going back to the platform is there you know, currently, is there a platform that you work with you work with multiple, like, I always ask people, you know, what are you using? And why did you make that choice?
Doug Shannon 09:50
Yeah, so we went through a number of different vendors to see the Hey, how do we want to do this? What are we looking for? There's never a good answer of this. perfect answer for everyone. Even when you get into a split with any kind of vendor, they say, hey, do it however you want to do it, which is great. I mean, it gives you the ability to kind of define it and an RPA space, you can do lots of different ways. No one has the perfect code, no one has the perfect, hey, do it this way every time. So we ended up actually going with UiPath. Over time, but anyway, mainly because of the fact that it was assumed to be fairly easy to use, they had low code solutions that they were developing at the time with either studio x and whatnot. Yeah. So we kind of went with that. And looking at orchestration, for example. So we always had the opportunity to do like power up on the Microsoft Store. And although it's fantastic, it doesn't have orchestration. So you end up having a lot of users having to maintain their own items, handle their own items, and it doesn't really talk to the bots or the processes. So you can have that mixed bag, it is automation, but it's not really RPA or Intelligent Automation. So when you're looking at democratization, when you're looking at building in like that potential hyper automation thing, we all talk about what we don't want to talk about. So that's where that kind of comes in. So you had to really, really answer that for us. And utilizing orchestration, utilizing the ability to capture the logs, create libraries, build those assets, really helped us define out how those bots are gonna be handled and how they interact with other processes for themselves.
Brent Sanders 11:29
Yeah, well, it sounds like, you know, you, again, you have an IT background, and you had mentioned kind of one of the earlier statements you feel like, that's really important. And I'm starting to think piece together. Why? Because it sounds like your practice resembles that of a typical software department, right? Or, you know, application department in, you know, in an enterprise. Is that sound, right?
Doug Shannon 11:53
Yeah, a little bit. Because, so when you have like an IT background, and so a lot of what I see a lot of groups doing is they come in and say, Hey, we're gonna have this business guy, come in and run their something, run their RPA, figure it out, like the business analyst, or project manager, go do this thing. And that's great. But without the background of it, that person may not know how to talk to these different groups like infrastructure, you know, your your back end desktop operations, like your your information security guys, so that they feel really more supported or trust that trust isn't there, really, because you don't know how to talk to them, you don't know how to say, Hey, here's why we don't need to do this massive security thing, because we have these other options. You don't have to, you don't have to talk about the technology so that they feel that you actually understand it. So then they're going to want to get more involved, because then go Whoa, what are these bots doing? What are these automation things doing? How do we handle that? What happens if there is an update? What happens if there's, you know, a patch, what do we do? So you really need to have that ability to have those conversations. And, and so that's why I would recommend, definitely you need somebody in it, or has it background, or has at least a good relationship with these folks. So that you can actually maintain that and build that trust, because not only you're building trust in it, because you know, RPA doesn't want to just be an IT tool. It wants to be everybody's tool, right? So you need to build the trust first so they understand, we're not doing crazy things. You know, the bots aren't going to take over the world. They're they're just doing what we tell them to do. And then you build that and you then build that trust and build on to the business trust and build on to the basic end user trust, so that everybody goes, Oh, this is for me, this is so I can have time back. I'm not sitting there watching. I just painted this wall and it's drying, I'm watching it, and this is really boring. So that's kind of where it goes. Right? So I think that's it.
Brent Sanders 13:39
Yeah, no, that's great.
Mark Percival 13:40
I was gonna say, I think that is something we hear repeatedly is which is the trust factor, right? It's like coming into an organization and saying, Hey, we're gonna automate all this stuff. And then a lot of times it gets, you know, sometimes you can get dropped on somebody's lap last minute, right, it finds out about it later. And I think that's an interesting piece with UiPath, which people are in Microsoft now. Right? They've been releasing more and more of these tools. I think you're gonna see more people kick off their own automations. Inside the company, they're not part of it. What are your thoughts on that? Because that's a piece I think, both exciting because you're seeing something solve their own problem, right. But then the other side of it is how do you kind of counter that with us? All the needs on it to actually have security there?
Doug Shannon 14:22
Yeah, it's like a three pronged approach on that one, right. So there's, like citizen development, that democratization in a way that's hopefully Intelligent Automation, once you get system developers in you, you may or may not see hyper automation, all of that stuff and like a whiz bang approach. So, yes, it's great to have anybody to be able to do that. I guess one risk that RPA teams will see is the fact that you know, since groups like UiPath, and some of these other bot groups out there, they have the ability to, you know, just download community Community Edition versions, so like, maybe they're already doing it. You don't know. Yeah. So that's when you get to work with your IT groups and say, Hey, man, install this community do that. And then they go search the network. And oh, hey, these people do like, Oh, hey, I'm gonna talk to those guys. So there's that, right. So you want to make sure you target the ones that are really interested in it. Because the biggest problem, I guess, I would say was citizen development doesn't mean No, it doesn't mean that this person knows how to do that. But as long as they have interest, then we can build upon that interest and show them that, hey, this is some cool stuff that you may be interested in. If you want to keep going great. Maybe you become a champion of RPA, maybe you become someone that we can actually install, the better, bigger version, because you know how to do code or you're already like a programming analyst and whatnot, and you really want to dig into it. Because you're like, you're really excited. That's great. But how do we do that? So you need a way to track it, right? Meaning the way to attract users and the way to, to build in that governance and say, here's how you work with the siwi. Here's like the Center of Excellence, right? Here's how you work with the Office of Intelligent Automation. So once I do think, before you go the system developer out, you need to ensure that all of your you know, i's are dotted and T's are crossed, because you need that governance, your own CEO, he needs to know how to do what they do, and why they do what they do. Because again, you're gonna be you know, explaining this out to them, you're going to be teaching and providing these lessons and saying, like, these are what we require of you to want to come play. And it's not like a pay to play, but it's more of like a just, hey, we need a basic understanding. So we know that that's going the right way. Yeah. UiPath has provided it has really provided but and you can buy it on mission hub. So tentative RPS pay to play for the other people trying to do the art, particularly using it Yeah, it's fine. So automation, how this cool because we're actually looking to start using that here pretty soon. And really, it gives us a way to kind of track like, who's doing what, and they can kind of enter in an idea. And I want to do this automated thing. And then you can kind of vote and ask some general questions about like, what's the ROI? What's the return on investment there? You know, where what groups are you attaching to? Do you want to tag different databases, systems and whatnot? What is that? how complex is that? So when that literally provides everything like theory, and he means so that we can share that we can address those questions, we can see is that value to us or the company? Maybe it's something we're not seeing, maybe we need to get another team to go work with these people to see what that value is. So that in that direction seems great. So that's the direction I would go in. And I'm planning on moving with this. But generally, to answer questions like hey, system developer, like, Are they a great thing? They're a scary thing. There just has to be like, held very close to the home and they say we will help you let's do this together kind of thing.
Mark Percival 18:00
Yeah. Yeah, I may it makes me go back to what you had mentioned, brought up agile, right? And like, that's another one where there's some parallels there. Right, which is sometimes you hear from teams like yeah, we're agile, and you're like, so you just have no process. Like, yeah, that's right.
Doug Shannon 18:15
We have to get out of our pants. Yeah. Yeah. Yeah, testing for the classic like they did the two, three rows like doing and done. Okay, keep going.
Brent Sanders 18:31
What is it again, this is like, I'm kind of asking you what you think, you know, it's not necessarily what you're doing. But what do you think is the right way to, to build in. So setting aside citizen developers, but like, is an ideal case to have an internal team in it, everyone building the bots we've had, and some background here, we've had a guest or two on that have really leveraged a outsourced model where they basically built a spreadsheet that hey, you put in your inputs in the spreadsheet will tell you if it's worthwhile to build this thing. And then all the risk is on this third party to manage to run in to meet an SLA. And they made a really good case for that. And what it sounds like in talking to you is that that's not how you're operating. So I was just curious, like, What's your thoughts on outsourcing the development and operations versus taking it internally? Like, what do you think you need to make that successful or does it make sense to go external?
Doug Shannon 19:33
Yeah, so I think the last team is fantastic guy. I've actually talked to him online. He does weird things where he outsource and I think for them that worked out great. Now, how I believe that things should kind of go you know, being like it like we talked about being kind of a hands on. I don't know how RPA teams learn how to do better. Or how to iterate better if they're not doing the work. So, so if you're, if you're going to automate something, that's great. It's going to fail. There's going to be some issues originally until you kind of work through those, you triage that and you understand why did that fail. And then you really understand it better. And then you can say, okay, we know how this works. We have like the PDD, for example, we can build like an STD, like a system design document, we can build all the other DVDs and everything else, right? So like how the business understands, like, what do I need to do to make sure I don't mess it up by changing the file or not putting it in the right place, that kind of deal. So you can build all that governance, and you can do all that. And the more governance you have the better ability, you have to say, okay, hey, run to a third party offshore team, can you go, you know, since our processes run like 24 hours a day, can you go watch these and if they fail, you can check back in these different documents to see what happened, how did it fail, and then go fix it, either run it manually, or have a tagging going until because of this, and then we can see it in the morning on our time to see like, hey, what was the problem here? But again, like, somebody's got to support it, somebody has to know how it works. Yeah, and I think so. So what we've done and what I've done and written recently doing so. So by doing, we noticed, is easy to kind of see how we can do better. And so instead of just having like a bot run a process, when the building, you know, some fault tolerance and like self healing and kind of ways into it, right? No, that's like taboo to say, but in some regards some groups. So, basically, you have like, what we've done is create all the bots, right? So all the bots can run any process at any time. So if a process is running, and there's an issue, currently, what happens is it may fail out, and that's fine, we have another bot spin up, and then go run that same thing, like 90% of time it works. So it's a matter of, you know, maybe it's the caching of the internet site. Maybe it's an issue with like, the VM or whatever needs to be restarted. And maybe it's an issue where something else we've defined and grayed out and said, Hey, what's the major issues that are causing this to fail? How to fix it? And so we just have another, you know, another bot come and do it, or it tries like a different login account? or what have you. So and if and we always like stop at a certain point, you know, we haven't done it three times, that gives us enough information to actually triage and what that looks like. So it kind of does like the footwork for us like why did it fail, instead of me going in there and looking at the logs? Well, I know that these three bots failed. And then to send us an email, hey, we fell three times, go look at this. And we have a lot of information at that point to go. Well, this one failed on this. This one the family chose SharePoint is horrible. Sometimes it doesn't want to connect and this other one failed, because they can't send an email because it can't find the thing. So then they go, Well, there's actually major issue here, let's go fix it. And then next time around like, it doesn't fail. Well, before we implemented this strategy, before, before that we, every morning, we had to check in. We're monitoring this and monitoring it Oh, that's still nuts on this, they haven't. But as soon as we implemented the strategy where things can run in tandem, or things can, can run if something fails, and then it goes and kicks off another process automatically. Not checking anymore with the monitoring is done. So in regards to the main thing I started this question was like, do we outsource? Do we do it in source? Or how does that look like? Well, I think I got them all. I don't need outsourcing at all. It just works. So it works great. And then yes, there's gonna be an issue with protein shake, once a week, this thing didn't run, why and sometimes the data doesn't run is just the self healing process where basically, it wasn't able to clear the cache, why don't care about that, you know, that that takes like, a few seconds to run. And then the next bot ran and ran and ran fine.
Brent Sanders 24:04
So I think if I'm understanding it correctly, you have a, I'm gonna use the term self healing this process to, I mean, it sounds like which we see this all the time, where you just run it again, oh, it's, it happens with my cable modem, it happens with my computer, just restart it, try it again. And for some reason for whatever it is, and by the way, I'm found this works for almost all pieces of technology. It's like you get into a state. And if you can just power it off and start it again. Everything works fine. And it's like that does capture a huge margin of errors of just Hey, let's try it again. So it sounds like that's working for you.
Doug Shannon 24:43
It works because I think it all started in the caveman era when we just hit it with a giant hammer and said like then we got to like cookies and stuff and antennas and like just come on.
Mark Percival 24:54
Yeah. Yeah, I think that's really interesting because I think a lot of times we You know, there's some, sometimes we try to squash all the bugs, we try to perfect every system. And I think you see this in every development, I remember doing it on the web development side, you would say, okay, we're gonna, like, get rid of every exception. And then we're and then if you didn't monitor carefully enough, you actually wound up in a state where you didn't have as much information as you'd like to have. So for example, you know, seeing all the requests come in, how many 200? How many for 404, that kind of thing, when you start seeing spikes in one direction, that don't kind of apply our Dart normal, then you know, something's out of whack. But when you kind of try to lower it to, like, I'm never gonna get in there. Unless like something catastrophic goes wrong, it's actually nice to see those errors as they pop up. But also to understand that they're not that's a normal part of that that's not a normal part of the technology. And if you think about it as that self healing, that retrying puts you in a state where you have more information, and you're not just going and tackling pointless errors when Yeah, you're right, they get resolved on their own. If you just retried it, you know once.
Doug Shannon 25:51
Yeah, that means you get those emails, 1012 a night and you're like, Oh, hey, it's all working. That's great. And you get the ones that didn't work like, Okay, and then you say, try again, alright, well, without again, but the third time it works. And then that's great. And you don't want to, you don't have to answer the email, you don't have to jump into it. If it doesn't work, you say, Hey, cool, I'll look at it in the morning, and there's no harm no foul affair, what happened? So there's that option, too. But like, that doesn't happen a lot, especially with how we're how we're doing this now. Yeah, it's neat stuff. But definitely, definitely, I think it's a good direction to go in. Because it's that fault tolerance, or we don't, we don't want to get stuck with stuff I don't, I don't want to be monitoring every single day, every hour of the day digging in, there was a lot of why did this fail. And maybe not even having the ability to hit those numbers or hit that, that bug fix. Because I can do more ROI. If I'm stuck, just, you know, written, you know, reviewing and monitoring where this is going.
Mark Percival 26:46
You had written an article on the self healing process, and you brought up a lot of things, but one of them I thought was interesting was just a lot of that went back to that trying to perfect the store, you know, trying to reset the state each time, right? Is that something in your practice? Like? Is it just I mean, how do you kind of do that in UiPath? Was it restarting the machines periodically? Is it restarting after every, you know, where do you kind of start that process as a developer?
Doug Shannon 27:10
Yeah, so. So basically, to start, like, the kind of like, the one of the basic things that we all run into, is, is just, you know, Chrome is very taxing on the system. Chrome is used by most people out there and most groups, so clearing the cache is fantastic. And that literally clears up a majority of issues that you're going to see because the bottom of me so fast, I only go to this website, do this thing. That works, I do another thing. And the next thing, you know, it breaks because it's like still thinking it's over here and you know, catch up. So when you have any process, or it's going into one automated tool, like processes, then it goes into another one. But it's the same thing. It just doesn't it doesn't know where to go. So to be clear, it works. And you know, you have to do like those triage boards, you have to dig into your code and see like, well, what is this failing all the time? And so when you do that, and you notice, like, what is that and then and then you build into, hey, let's restart these VMs every once in a while. Or let's even look at I think, as of yesterday, we're looking at Oh, hey, what's on these Windows updates, like hanging out on the VM that needs to be cleaned up? Let's go fix that. How do we do that? And then basically, we're gonna have a very long kickoff and information like, Hey, I'm running out of space. Hey, UiPath orchestrator, we're doing that like, let's go fix that. And so the bottom of it, we're probably gonna have it to where the bot goes and clears out space, you know, run by the basically delete temporary files, does it all on its own, and then does a restart, and it comes back up and says, Hey, I'm back up and work. So that's, that's probably, you know, another step we're gonna go in. But again, like you don't want to do, I've talked to some people online, especially from the article, right? And they're like, oh, we're just gonna go crazy, and it's never gonna stop. It's like, well, that's why you're an orchestrator. brains of the thing. And whether or not the bots are dumb terminals that do whatever they're scripted to do. But the orchestrator has like the thought process to go and it's not AI. It's just, you, as a human have to be the eye and go, when do we stop this? Okay, we run it three times when we stop. And then on the third time, we say, Hey, we stopped, it wasn't still faulted. If it was send an email to the CSV and say, Hey, I'm having issues here. Go check it out. So we just have to think outside of our own box, right? Like, yes. You know, we have these automated processes. Yes, they can do some cool things. If they are automated incorrectly, they can do some bad things. I mean, we can talk about how like, people are buying up all the xboxes and PlayStations and say, Hey, we did it was great. Like, that's a bad thing.
Mark Percival 29:44
Yeah, I mean, they interesting by me going into the self healing, a lot of that's just hard won knowledge, right. It's just you had to kind of go through the pain points. I guess that's the hardest part about this base sometimes is that getting into it, it's so overwhelming at first until you kind of get enough practice going and understanding all these issues. You just it's a it's sometimes it's a big hill to climb.
Doug Shannon 30:03
Yeah. And that's kind of what you need it background I think I think I think having groups that come in that have it background myself and and other people I've talked to, they don't have as hard of an issue of doing the triage of doing the, why did this fail? And like the analysis of like for the sale, but the why and how and how do we fix that. And then what's the future look of that, because you have to have that feature look, or you're going to get tied down in some other process or other groups that have hundreds of bots. And then they end up just having one process per one button when we have like 10 bucks, and we are saving tons of money. And we are doing this and we actually have more bandwidth out of our box. By having them be able to do each other's work or having them do the other one does that up right now this one can just go run it. So the feasibility and the usability is very fluid. And going that direction was a fantastic choice. Mainly my part of my team's part. However, we wouldn't have known that going back to the original version, we wouldn't have known that at all, if we weren't in there doing the work. So if we were to outsource that, we'd have never seen that ability to do that we would have only been focused on like, well, what's the next thing? Is there value? How long does that take. And really, the rigor needed, and the governance needed to be able to hand something off to somebody is much longer, I think the average time that other people are talking about it takes like four to six weeks for animating insanely long. So just an example. So our normal cadence for our Sprint's on me having an agile background, and then Scrum Master whatnot. I mean, of course, I want to make sure it's very iterative. But we're doing an eight day sprint. So eight days we do it, eight days we release done. And we're doing a good amount of effort, we're usually you know, one or two things that we're getting out that are automated. And while maybe to like you know, critical paths and whatnot, we don't want to pass something out and say, well, you have three or four tires on your car, go ahead and go, like we don't want to wait to kind of hold it back here and there. But we move very fast. And so I think the reason why we move fast is because we have the ability to use the box and the automated process in this manner. And then we're not stuck monitoring, we're not stuck in dealing in like the murk of really like, oh, there's this thing's running, and I don't know what they're doing. We're in it. And we know how to fix it. And so I think eventually, we may look to a model where we do outsource, but it's going to be we have a lot more time. Now. We don't have to do that. Because we have this self healing these fault tolerance stuff in place. So we're not forced to go down that road because we have to. Some groups have to because they can't support it. So if you can't afford it, well, then you're adding more money, which then decreases your ROI and decreases like the value of what you're doing. Yes. We're also doing valuable things until you automate anything, there's some value. But is that value increasing? Or is your cost increasing? Or are you paying someone else to do whatever metrics eating away at your bottom line?
Brent Sanders 33:05
Right, right. Yes, feeding the beast at that point. I mean, that's, that's a horror story that, you know, we've heard, you know, undisclosed company, but a company we spoke with that has been in the RPA space for a couple years now. But they've finally hit that tipping point where Sure, they were able to make massive improvements on headcount in one area, but they had to add, add a bunch of, you know, high cost it hires in order to keep the bots running. So you always have to keep an eye on, you know, where's this headed? It kind of that's, that's my next one. I mean, not necessarily about headcount. But I'm wondering, like, how do you see your practice shifting over the next five years, like, let's just assume, you know, the rate of automation slightly increases. I should throw a little background in here, Mark, and I have always had this kind of interesting, or interest in legacy software. Right. And that's actually kind of what got us into RPA was, you know, its application with legacy software. But do you feel like automations that, you know, let's say you built last year, and that you don't have to touch and they're always working and fast forward five years? Like, are they going to be the new legacy software? Or how do you see it scaling? Like the whole enterprise, you know, with, let's say, hundreds of automations in five years? How do you keep track of all of it? How do you think it's gonna go?
Doug Shannon 34:34
Yeah, that's a hard pill, the harder is five years, a long time, and it requires a lot of time and RPA. Right. So RPA is quick. And I mean, I think I think, again, governance is key. how you handle that, how you maintain that, how you explain that, how you build in, you know, the flow, or it's like a flow of what they're used to getting in the flow of the ideas that are coming in, like what do we do? How do we do it? I do think that once the business and other teams and whatnot, just talking generally, once they have more ideas and understanding about how it works, that that they'll come into different ideas where we can, we can actually be a cost a cost center where we actually increase the cost or sell out services even to even outsourced in a way. So this is the ability for RPA. You can, you can do different tenants different sides. And so you can literally split your information up in a way that where you can do even different companies. So like, if a company came out of the works or some startup and they wanted to do a thing, they can literally have everyone sitting on their main orchestration, automated governance still split it out. So there's lots of ways to do different things. I think even over a UiPath they had their rebook work, we built work festival of the last three days. And like one of the things that their chief evangelist, Dr. Kirk was bringing up here, like six different Hey, or five different things like this is the future. And I think I think this kind of goes into kind of where I think as well. So he was talking about, you know, RPA, the new GRP. We're in the 90s, everybody's a European, we do stuff and it's all great. And then RPA, is they come in and say, Hey, here's how we tie those API's. Together. Here's how we take that legacy stuff, and make it new, because you can't talk to it. It's very old, it doesn't want to work. So here's how he builds integrations. So RPA can definitely the bread and butter, right? Like it always works in that area where it's like we make stuff that can talk to old stuff, new again. So that's one example. So he does, RPS DRP and the other one is in the automation series will make money and I'm seeing that right now is where teams that understand RPA have the ability to say, Well, hey, we have this thing coming up, we want to do this, this new thing. And we want the new contract with this with this entity. So we know before we've done this, and it took like 10 to 20 people, how do we do that better. And so we actually were able to work with a team like that and said, Well, you've never done this, you have some experience in the past, like a different kind of team mechanic. But now you want to do this in this fashion. So you know what's needed. You know how much it costs, then you know, if you have to outsource, and they have another cost as well. So that's great, we have some information to understand the ROI we have, some of them are astonished at what we're getting ourselves into. So we ended up doing instead of taking those 10 people to do that job, which was very manual and very like, hey, do this thing, Hey, get this thing back. We ended up doing it. And we only have two people managing that team, everything else is done by robots or automated processes, right? So that's a huge savings. And so then the business goes, wait, we can do stuff like that. But yes, we can be a center that actually makes money, like we're actually making money and how versus not going that route, and then kind of stuck with it. So that's another example like he, Guy Kirkwoods predicting the record and doing it. Another one is like his, his prediction right at the islands of automation will be needed, will need to be bridged, right. So, automation being you know, what, if you're doing SAP and you have AI RPA, what if you're doing blue prism and automation anyway, a lot of a lot of groups and teams out there are doing one or two or three. I don't know if I'd recommend it. But if you do it, you get those islands of automation, and you have to put them together. I know. We talked about Microsoft earlier. But Microsoft power automate is a great tool that you can even give us as developers you can give to individuals that just want to automate things. And then maybe you bring that into UiPath. And I know from talking to them, they're looking at building those integrations and those activities to make it easier because they want to be that center of Hey, bring all your stuff to us. You know, here's what we can help, here's what we can integrate. And so that that's great. So if we can when we can use UiPath. And then yes, they should go public because they're gonna be doing some neat things to get people moving away from other maybe nonstandard or other things that aren't fully RPA to be RPA. So, four and five, four was like, you know, robots will work from home robots will be created. Well, it's kind of already being done. Thanks to COVID. We're all working from home, we're all working remote in some way, shape, or form. And although that's great, and it's moved a lot of companies to go that route, there's also a lot of potential loss there for smaller companies, why not because of the business and whatnot. So but work from home robots. I think that just goes back into UiPath, for example, where they want like a robot for every person and although that's a great idea, it's gonna take a lot to get there because that's kind of goes back to that hyper automation ideal where the, Hey, everybody has a bot, everybody can do this thing. So, you know, I type those key types of key and then it learns with me and then I go cool. He learned that, go do that now. That's all great and It's not there yet. It's almost there. Or at least like we're almost there, we're trying to get there. Over the last one, just for example, he was saying that employee experiences will be an important as customer experiences where, yeah, it goes back to how we started this. So you gotta teach, we got to teach everybody how this works, we have to make sure that there's a general understanding. I know from the community out there of RP professionals, something we always talk about, we say, hey, why do we say hi proud nation that kind of scares people? Why do we say democratization of information or, or Intelligent Automation and automation? Okay, I always like terms, like, Why can we just call it one thing? Or why can't vendors stop calling different names and maybe just stick with one and say, let's all agree to call it this one thing, because we're all gonna win if we actually have a conversation. And we're all talking about the same thing, because the more we mix up our end users in our business, you know, the CIO is out there. And those guys that hold all the money and say, Hey, we want to, we want to save money, we don't want to spend money. And they're they're asking the questions, and they're reading the books, and they're getting mixed up. But they, but they want to do the thing, but they don't know how to actually ask for it. So I think that kind of covers those silly points from him. So especially I bring it up.
Brent Sanders 41:14
Yeah, no, I remember seeing the RPA is the new GRP. I saw that yesterday in some posts, and I was like, that kind of nails, kind of nails in what you said about the the islands, I mean, I think there are so many products out there right now, like, I feel like this 2020 has been the year we'll see of 2020 has more, but I've seen so many different platforms, even, you know, we've had a couple of them on our show where it's, you know, they're small companies that they have their own platform, and everyone has some version of something. And, you know, our prediction has been that they're going to, I mean, use the word democratize, they're all going to get really cheap. And I think you're gonna continue to see, accounting went with automation anywhere, you know, another party HR, they bought into some other automation platform, and everyone's seeing the value in it, but no one's in big companies, it's hard to coordinate everybody. And so that's where, you know, the some companies have that governance around what they're using others, you know, they're inverted, the, you know, it is in service of all of the other parts of the engine versus it kind of leading that way. So it'll be interesting to see how it all kind of shakes out. But you know, it's definitely a good set of sort of what's, what's to come from guy, we had him on the show, and he was always full of interesting future perspectives.
Mark Percival 42:39
Now, this is really interesting, I think, you know, one thing we don't talk enough about is sort of when these things kind of fall off the rails and how you kind of handle this stuff. And I think one of the things that we always see as well is, is certainly this idea of imperfect state for bots, right? It's just a really common problem that you run into. And it's interesting to see, it'll be interesting, see, where the, the, you know, where this industry leads, because I think you're you're seeing, like I said, it's a lot of hard fought knowledge to get to this point. But it's gonna be interesting to see how UiPath and the rest kind of push this out and find better ways to kind of encapsulate these things and, you know, less, less surface area of failure, right, like less things that can go wrong. Because there's a lot of stuff that's just it's, you know, it's complicated. So, ups, ups the failure rate?
Doug Shannon 43:25
Yeah, there's definitely a lot to talk about. And then also, just like the IT knowledge, right, is that the team is really, yeah, whatever team you bring in, make sure you like them can talk to them, and they can actually, you know, get in front of leaders and explain what's going on. If you don't have them explaining it, I mean, a lot, a lot of posts, and a lot of people talk about, Hey, you know, you got to get the leaders involved, you have to get the leadership, also, they understand what value you're bringing, because if they cut, you know, cut the first time and say, Hey, you can't do anymore within what do you do with what you had? Well, okay, we have other stuff automated, or do we just stop, like, you're gonna, you're gonna have a lot of problems. We're not gonna, we're gonna start picking that up again, you're gonna need to hire bodies and people and this, and they're not gonna want to do that nobody wants to do these things. Automated. So, eventually, I see definitely how the future goes, especially the next five years of people. It's a very long time. And then, in regards to, you know, democratization and hyper automation, all that stuff is all interesting. And it'd be cool to get to, I do think a lot of people are waiting for the whole AI ml movement will lead to machine learning in AI, and artificial intelligence, and then we'll be okay. We don't know that we have no idea that we are seeing some of that now. I know peg is doing some cool stuff with AI. I know. prisons doing some interesting email stuff. So but nobody has the right mix still. So it's a matter of since we've had so many different tools and different people doing it. Like I said, like there's no there's no right or wrong answer, right. So there's no, here's your playbook. Here's how you get it done, right? Here's how you save millions of dollars in the first year, like, Sure, I could write you lots of things and how we did it, but it doesn't. It's not gonna be right for you, right? It's fine for everyone. But I do think that you need to get in there and do the groundwork understand what happens, why you know, what levers you need to pull, and why they fail. And then by understanding that, and doing triage, and you know, using agile methods, you know, really give it in, like, look at his prayer reviews, like, how did this fail? How did we do? What are burned all that stuff? it all plays a part. And I didn't think originally agile would play a part in RPA. But really, it does, because you are building iterative things, you are building something to get it into production. And it really works together. So that's sort of what I was planning on writing about is basically Hey, what's next? Like, well, do it agile wise? Because if you do, you're gonna get more value and really add on all that value. So an arcade is trying to build that ROI.
Brent Sanders 45:57
Yeah, yeah, that's a really good way of putting it. I mean, it's, it is a great process. I mean, still apply it for software engineering projects. I mean, it still works incredibly well. And, you know, I don't know of anybody who's doing any form of, we were talking about this on an earlier conversation, but I really don't know if anybody's doing it, you know, the old school sort of waterfall way, just because the ROI, even on the process to gather requirements will will start to invert, you won't really see, you know, it's like, hey, I need these things done. Let's chip that off. Let's deliver the value of the highest amount of value in the shortest amount of time. Let's work on that first, and see what happens. And it is difficult, you know, for that sort of uncertainty, will we get something out that has that value? Is there enough to ship? And I think that's the fun part about RPA. In automation, in general, there usually is it's usually like, hey, if we chip off eight days of work, we're gonna deliver some value, and we'll be able to keep revving that engine.
Doug Shannon 46:56
Yeah, definitely. I mean, that's, again, where we have to you have to go to explain because a lot of stuff that may come in, you get in a waterfall type of format, and what we wanted to do this and that thing, and we want it done here, like that. That's great. How about if I give you a little bit up to that, and then we say, we'll turn it all on and production when you're ready? And so it's a conversation and be like, yes, we're going to literally provide pieces of what you want, and really just explained that out again. So it's all, it's a lot of teaching. But in the end, it works out pretty well.
Brent Sanders 47:27
Yeah. And it for us, I feel like you know, we have used that process for a long time. And it, it feels really natural to Hey, let's let's thin slice this huge process that seems really daunting at first. And in our experience, it's like even chipping those first little dribs and drabs. It's like, oh, wow, like you already started seeing the time increases, like people getting their time back. Or, you know, honestly, we're seeing more projects that and we've, I think mentioned this before, it's not so much like we're saving headcount, it's the like, this just wouldn't be feasible for a human to tackle, right? It's like we're going through 1000s of records or something. It's just like, you can't even do the ROI math. Because you wouldn't even do this with a person. It's like not, it doesn't make sense. So it's like going back to an abacus versus using Excel. It's like hard to understand how somebody would even do it.
Doug Shannon 48:20
Yeah, sure, reminded me of a question. So. So I see a lot of RPA teams having issues defining ROI, right. So you have those situations where, where you are, it's really hard to define. So a lot of teams are going well, I'm working on this specific automation, this person makes as much a year they're saving this much time. So then we'll do the calculation. Well, that's, that's really hard. And it takes a lot of time to go, Well, how much do you make two years plus know how much we can talk about, but we don't need to? So I mean, and then you get jealous, you're like, Why? What's going on? So? So to avoid that? What we did is we talked to our finance guys and said, hey, what, what is? What's that number? What's that number that everybody? Like? is around? What's the average? So you just give us something that says Like, this is the overall cost of an employee or a person. And it's different between different organizations, different big companies, they had like three different numbers that they gave us what we said, we'll just take the lowest number and run with that because we want to, you know, see value, and not just pump up numbers and pump them up or not. We want to provide like, Hey, what's the general so anywhere from like, the average person in general is like, you know, 540 to 50, or even 80th are really awesome, but so they just go with middle and mid range. So hey, we'll take like 6052 Hey, average person benefits, whatever makes like around this range. And then we use that for everyone. Yeah, and then we say, okay, cool, how much time you save, always have two hours per week, from what whatever. And there's cool times and then they go oh, that's really easy. Yeah, it's just one number, okay.
Mark Percival 50:03
I like the idea of you going to someone and saying, Hey, I'm automating your job, I need to know how much you make.
Doug Shannon 50:10
Nobody likes that conversation, right? What do I do? Like we're gonna do more? Well?
Brent Sanders 50:22
Yeah, I mean, it's it. Like, we have that calculator on our site, right? This like self assessment, and it's only part of the picture, because it's too hard in a, without having conversation with someone to say, Hey, this is what it could be worth to you. But it's like the baseline consultants, you know, here's what, based on the numbers what we would be saying for saving for a given amount of someone's time. And there's always this thing, I was like, Okay, well, you're just gonna make everyone's lunch break longer. And it's like, that wouldn't be a bad thing. I honestly, going back to what Guy was saying, it's like employee experiences are really important. I mean, especially in, you know, we've seen customer service contacts, and I think pega pitches this in their marketing materials, but like this use case of, you know, a call center employee that can either spend most of their time filing, you know, filling out forms and making people wait on hold and dealing with cranky people, or being sort of this wizard that can kick off processes and say, Hey, I filed that for you. We're all set, and move on to the next call and feel far more productive. It's like that actually has a strong intrinsic value. But how do you put $1 value on that? It's like, I have no idea.
Doug Shannon 51:30
Yeah, I mean, that's, yeah.
Mark Percival 51:32
That's like, I think I think, in general, if you see the automation, and it's easy to automate a process, it's easy to automate, it's probably a horrible job that no one wants to do. Right? Yeah. Right. Like, it's, it's usually never that fun. It's like, if it's easy to automate, then it's pretty like boring, horrible job that somebody would be happy to give up. And then you do give them a lot more freedom and autonomy to kind of do something that's more valuable and more fulfilling? Yeah.
Doug Shannon 51:54
I think that some of that controlling that drives, I've always believed, especially seeing how it has changed over the years, how we went from there, IT centric and it kind of driving business to where business is now driving it. And then now we offshore, a lot of it. And then especially here in the Bay Area, whatnot. So you know, when I was actually working at Cisco, we do a lot of like acquisitions. And so seeing how they started to work. A lot of people are driven to startups, because of the way that startups are handled, they like the ideation, they like to be able to innovate, like, everyone there is very happy, even though they they make nothing or make some or the hope and my belief that they're going to be acquired or built into this greater thing. That it's like a driving force. And it's amazing to see the amount of energy everyone hasn't like, in the startup area. So what is that, and I've always thought, you know, it needs to, to get back to doing customer service, it needs to be able to handle and handhold and walk people through things because it can be scary. You know, you have your older age group that doesn't understand technology, you have ones that do understand technology, but then there's still like a large gap and even people come into, there're kids now and then talk to you. And they're like, well, we don't know anything about this, like how do you not in the Netherlands, or like I only focus on Instagram? And there's another thing? Like, that's cool, but what about this? So we're always going to need customer service. And I think I think that's somewhere where we're lacking as a culture is like, Where is that customer service? Because when you even when you offer something, for example, you lose a little bit of that. camaraderie, customer service is usually like, Oh, hey, do you understand what I'm saying? I understand. Are you saying are we on the same same page? Like where's that connection. And then once you get that, it's great, you know, you understand them, they understand us, and we all work out to make it better. But there's there's something to be said about customer service and providing time back especially to those people on the phone, they just want to get that cell because when you build that that relationship, you're going to get more sales and you're going to get reoccurring sales and that kind of goes back into like quality assurance, right? It's you know, if you build quality like Apple did you know the whole like the golden rules of selling if you build quality and you build it well and as a good presentation, people are gonna buy it because they like it they understand it easy to use all that stuff and you know, Apple when huge instead of going to the Microsoft market wherever then you can just build your own but what Microsoft did in a way where people had to understand that nobody understand what at x and knowledge boards were and PCI and stuff right cool those who made great systems and they were fast. We all play gaming off of it who's fantastic. But then the rest of the world said we don't want that we don't understand that we're gonna get this simple thing that looks pretty and it worked well. Hardware driven stuff is amazing.
Brent Sanders 54:51
Excellent. Doug, any any parting thoughts, anything you wanted to leave behind with our listeners who want to promote your SoundCloud, something like that? Now's your chance.
Doug Shannon 55:06
Yeah. I'm one of those new users right, understand this. So, yeah, I think I think our space, I appreciate all the people out there, the LinkedIn world and how we try to connect and whatnot, I tried to provide information back that I learned, because I think that's the value, right? So not only am I teaching my teams and people I work with and our own business and people, other people I previously worked with, it's a matter of teaching. Well, here's what we do. And maybe we do it wrong. Maybe maybe self healing automation or fault tolerance is the wrong way to go. Maybe having bots that can run any process, or the ability to, to have it to where you have one bot has one access, because it uses one link to go to a certain account that has permissions rules. And then another bot goes, Oh, well, I'm going to finish out this same automation. But I'm coming in from different angles with a different user account, to build in that honorable action item of separation of church and state in a way where you don't have one bot doing everything, because that's a bad idea. And then there's also security risks, right? So I think, just I'm glad to be a part of the community. And I'm glad to work with everybody online. And I think it's a great community. And I think that, together, we'll all get to a point where we push for that, that nomenclature of this is what it's called, and this is how we do it. And here's, here's some good tips and tricks, right?
Brent Sanders 56:27
Yeah, well, thank you so much for coming on. And I think you know, as a, it's a great way to contribute back to the community, get on the show, and like, tell us how you do it. I mean, your articles are great. And keep that up. I mean, our mission is, is kind of part of that is like trying to put the word out from the actual practitioners. And there's all this sort of marketing junk, the white papers, the magic quadrants, and, you know, hearing it directly from practitioners is what we're trying to be all about. And it's been awesome to, you know, have people all over the country all over the world, I should say, come on the show talk to us, as well on LinkedIn, which seems to be like the, probably one of the better places to, to, you know, have these conversations and find these actual practitioners and share, you know, thoughts and ideas. And that's, that's been to me, like the most eye opening thing it's like, and I don't know why I don't get really excited when it is global. And that is the internet. I mean, it's nothing new. But it makes it that much cooler, right? It's like we're, you know, through automation, being able to kind of get all these different perspectives, meet all these different people. And so yeah, I can only echo your sentiment around how great the community is, is evolving into.
Doug Shannon 57:42
Yeah, it definitely seems from the community out there that the European community is almost leaps and bounds ahead of what we're doing here in the United States in a way because just talking to some of those guys, like wow, like you get it you, you get in your users get it? What did you guys do? in the water, so it just got to something right. So maybe it's just their accident. They just like the bad guys for survival or whatever, like, Man, you really get it like it's amazing. So yeah, so cool folks out there and appreciate base help, and even learning stuff on the way but I'm trying to be a part of it, where I can provide what I've done. But again, I wouldn't be able to done what I'm done without really getting into the weeds and really understanding how the things work, to where we build out and really some interesting stuff.
Brent Sanders 58:35
Well, Doug, thanks again for taking the time to spend with us and talk to our audience. And yeah, thanks so much for listening to our audience.
Doug Shannon 58:44
Yeah, Mark, Brent, I appreciate it. Yeah.
Mark Percival 58:47
Thanks again, thanks Doug.