Getting Started With Robotic Process Automation
“According to a recent report from Grand View Research, it is estimated that the RPA market, which was valued globally at $125.2 million in 2015, will grow to be $8.75 billion by 2024.”
– Grand View Research
While the technology is not new, RPA has become the hottest topic in technology over the last several years. It’s not surprising that many organizations are struggling to find a starting point that will drive the highest ROI given the vastness of tasks and processes that need to be automated. Additionally, how should organizations think about the many players in the space and align their investment with the right technology partner?
Join Francis Carden, VP of Robotics & Transformation, and Ying Chen, Head of Product Marketing, Pega Platform, for a practical conversation on these and many more questions including:
- What is RPA and how is it used?
- Why the buzz around RPA when the technology is NOT new?
- How to drive the fastest ROI for RPA?
- How to drive new and recurring value into existing RPA investments?
- How to sift through the multitude of options from the market?
Learn more about how Pega Robotic Process Automation (RPA) can help you eliminate the need for human intervention in automated processes and free up your employees to perform higher value work.
Ying Chen: Thank you for joining us today. My name is Ying Chen and I am the Head of Product Marketing of the Pega 7 Platform.
Francis Carden: Hi, I'm Francis Carden, I'm VP of Robotics and Transformation.
Ying Chen: So we know there are a lot of questions out there about robotic process automation. We know that whether you're starting your journey or you've already implemented RPA, there are so many use cases out there, and there are lots of vendors to choose from. So what we wanted to do today is to share some best practices and some observations on RPA and hopefully we can answer some of those questions that you have.
So just to level set, what is robotic process automation? So essentially it's the use of software robots that mimic the work of humans across the application in a non-invasive way. And what that means is to leverage the robotic automation, you're not making any modifications to the underlying technology.
RPA is really, really great and helpful for tasks that are repetitive, that are also high volume, who are also long running. And for the most part, they're being used in the context of tasks or processes that are really rules-driven, meaning you don't really require any human intervention as a part of that process.
It's also estimated by Grand View Research that essentially the RPA market place is so hot that it's going to reach $8.75 billion by 2024, as well as the fact that it has so much cost saving potential, where it essentially costs 65% less than hiring a full time employee.
So from a technology standpoint, automation is not really a new concept. Right? That the concept of screen scraping, using scripts, as well as BPM as a technology to automate task that are highly repetitive, that are rules-driven, doesn't seem to be a new one. Why with your 30-plus years in the industry, why the buzz now?
Francis Carden: Well, I think, Ying, it's really easy. If you look at it, it's called robotics. What's not to like about that? I think the market has really taken that whole AI and robotic theme and applied it some of the older technologies. And actually that word, that RPA, is now being used for pretty much anything that can be automation. And you know, there's nothing wrong with that, right? I mean it's a good theme, it's got the C-suite excited. Everybody's looking for that nirvana, that next generation and way to actually solve some of the problems we have today in technology.
Ying Chen: Interesting.
Francis Carden: You know, if you look at it from the perspective of who isn't talking about it today, there's almost nobody not talking about it. The direction for cost savings and quick wins, that's been around, I mean, I've been doing this now for 30 plus years. And I think my first programming of a function key was really my first robotic automation, really, as we would it today.
But everybody's looking for that quick win, right? I mean, cost reductions are coming. Everybody's trying to drive that into the enterprise. And so, whilst they're on transformation journeys, which most organizations, if they're not on, they should be on. But going back to looking at our biggest cost, which is human capital, this is really what's driving this need for robotics today.
Ying Chen: So going back a little bit about the application of the word robotic to this technology. What's your view in terms of how smart these robots actually are?
Francis Carden: Well, initially, they're not very smart at all. They're actually taking the rote work that humans do. So literally training a human and paying a human to take five fields from one screen and putting them into another screen. Makes no sense. Okay? So in an ideal world you do that through integration. You resolve that problem and you would have one screen on your desk. And the reality that's not what it's like today.
So robotics is really designed initially for taking all that rote work and just moving it away, moving it back into the computer to stop paying the human to actually do it.
Ying Chen: Makes sense.
Francis Carden: And it impacts, I think, a number of other things as well. Right? You know, it's not just cost savings. There's a lot of rules and governance and compliance coming up. You know, that the fines are getting greater. I heard that some of the fines actually outweighs massively, even said to the point where it's a percentage of revenue now.
And so, humans do make errors, even for simple things. Move these five fields from this screen to that screen, but you move the wrong field once, it could have a major impact on your business. And so, you know, having a robot do those steps as opposed to having that human do the steps means that you're eliminating and reducing the amount of times that error is going to occur. And the governance has been driven into financial services right now. Is driving a great push towards robotics as well.
Ying Chen: So that's really interesting. So you're talking about use case where these robots are essentially being leveraged to be able to consistently apply almost rules to these processes. Where, you know, humans might commit errors. Is that kind of what you're saying?
Francis Carden: Yeah, I mean, we just do. I mean, every day. You send an email to the wrong person because it's just as easy to drag over or not hit the copy when you meant to hit the copy before the paste. So it just happens, right?
Actually, we're much better at doing that cognitive work where we need to think about the process or do we need to insert an extra step of security because we've got a bad feeling about it. Humans are great at that, but they shouldn't be making decisions around stuff that's very clearly rules-based. It's just a perfect fit for robotics, automation as it really ultimately comes down to.
Ying Chen: Gotcha. So another question that we get a lot when we first meet with customers is many customers who have already implemented RPA are asking themselves the question "You know, I'm three years, maybe I'm five years in with my RPA investment. Somehow I didn't actually get to achieve all of the outcomes that I wanted to get to." We've spoken to one particular customer recently about the fact that they're three years in within their journey, and they were able to only leverage about 10% of the RPA investments.
So I guess where's the beef, right? All of these fancy promises, including the fancy market share and the cost savings. Where are we right now? Why isn't the technology delivering?
Francis Carden: Well, so there's a couple of things. Firstly, if you look at the reason for automation is because there is just too much manual work, right? And over the years, we've been looking at trying to actually replace that manual work. We've got lots of screens, right? And so, we constantly train the human to be the robot. Right? To be the automation engine, and it's an expensive way to do that.
In the perfect world, you would just replace those applications. You would re-engineer them. You wouldn't require that if you like those multiple applications on the screen. But it's just not been easy to do. It's very expensive to re-engineer a process that might have 30 years of legacy code behind it. And that's 30 elapsed years. I mean, that's thousands or tens of thousands of man years of code. And somebody would have to wake up one day and says "I know what that process is supposed to do. I'll re engineer it." It's not going to happen overnight.
However, back to your point about why is it not as successful all the time. There are places where it's very successful. So there are places where it absolutely will deliver value. But we're seeing that the analysts are starting to say "Well, hold on a minute. It's not delivering all the value all the time across, you know, at scale." And one of the reasons for that is because there's not much easy stuff left. Most of the processes that we're paying humans to do are quite complex. You know, I use the term that there's more branches across the process. It's not just like do this, do this and this and this. It's like somewhere in the process is like well, you could potentially go off in a thousand different directions.
And so, if you try to automate a process that looks simple because you're thinking well, the humans not paid very much or that's a very straightforward process. When you actually come to map that process out, because they're not normally mapped even though businesses like to think there's a mapped process, you realize is actually, you know, potentially could take three, six, nine months to complete that process. And it might be half an FDE or two FDEs.
So I think it's very important about how you go about doing robotics. Talk to a lot of the organizations out there that have got experience with it. But don't just look at the successful ones. There's a lot out there that has not seen the success. In fact, there's a lot more who have not seen the success of robotics on its own because they're looking at one particular type of technology. So I think it's around trying to deliver, and they have high expectations because it sounds good, but not doing the things properly when you actually go to deploy it.
Ying Chen: So coming back to that point, I guess I'm a little bit confused, because if you look at something like claims processing, as an example, where it's very much rules-driven and previously tedious from a human intervention perspective. So you figure that by now, given all the technologies that are available, including RPA, we would get to 100% efficiency. Right?
And so, when you kind of talk about the focus on the use case, and not just kind of only focus on these things that can be 100% automated, I guess tell us a little bit more about where are these additional efficiencies that you can gain outside of these fully automated end-to-end process that don't require any intervention.
Francis Carden: So, if you look at claims processing, it's a great example. Let's just say, you know, I think there's a few years ago survey done where somewhere between 88 and 95% of claims processing is automated. Auto adjudicated as term I think in healthcare it's called.
Ying Chen: But why isn't that 100%, right?
Francis Carden: Because it starts to get more difficult as you get... You know, we can easily look at a process and say "Well, there are steps in that process that's really easy. Move these five fields from here to here." But what if this and that, and this and that? And that starts to become where, you know, even Jane, who's been with the company for five years, might turn to John and say "John, can you help me with this?" But that's really difficult to train the robot to do, okay? Unless every single scenario is mapped into that robot, it simply can't do it.
But even then, in the last few years, robots has been used to get another two, four, five percent. So we're getting there. We're at 95-98 and sometimes 100% in some of those processes. But now your cost of return, because you've got less people doing that now, right? Still a lot of people, but each five FTEs for full automation, to automate an entire process, you might only have two people doing that exception, five people doing that exception, or maybe 10 at most. And a lot of the robotic opportunities that are being delivered are around smaller groups of users in that market.
Ying Chen: So can you give us a little example. I mean, here we're looking at a picture of Kerim, who's our Head of Product at PegaWorld. And I think he was talking about the address changes that were anticipated going to Canada. So let's talk a little bit about why that is a non-traditional use case, if you will, from an RPA perspective.
Francis Carden: Well, so if you look at this screen, actually, what we're trying to also show is it's not just the computer. I mean, 30 years ago, who would have thought that we'd still need paper. Right? Because when you're on the phone to a customer or you're trying to actually assimilate some data that's may be coming from multiple points. I actually had somebody once tell me it's easy to write it down than actually go to the thought screen and forget what I gotta do.
So there's a lot of that manual work that's still involved, you know. So the systems are not only disjointed because they were written in different technologies over different times, but we've had to spend our money training that worker to do all those processes in between. And it's just ripe for error.
And so, just a simple process of moving, it's not just moving and changing the address, it's maybe going to my provisioning system and saying "We need to ship some new equipment. We don't want to ship it to the wrong address." There may be legal regulations when you move out of state to another state. Insurance regulations are going to change. It just snowballs in terms of the variances that go on. And that's why you see these desktops just full of these different applications because it's just going to take time. It's going to continue to take time to re-engineer those processes so that you don't need to use robotics to do it.
Ying Chen:So that's really interesting because it sounds like what you're saying is that there are still places that robots can't do everything that we can do? Or is it true that they actually can do everything that we do?
Francis Carden: I think it's clearly that robots cannot do everything. I mean, I think most people who's looking at RPA or robotic process automation, they realize it's not applicable everywhere. The question is, is it 1%? Is it half a percent? Or is it 20%? And so, that's why it's really important to step back and say "Okay, what is it I'm looking for? I'm looking for an ROI, where I can have less people doing the same work? Or I'm trying to grow my business, and so therefore, I'm going to be taken on more work. But I don't want to double the size of my organization?"
And so, what the RPA vendors, when the term definition RPA came out to replace all of that terminology we just talked about. So RPA, robotic process automation, a lot of the vendors only had one way of doing it, which was to automate nearly all of the process or everything else is done by a human. Well, when we started out in the business, we said "Well, actually, if you look at large groups of people, say in a contact center or even in a back office, they're doing many of the same things. So if you could automate the many of the same things across a large group of workers, whilst those workers are still there, you've effectively applied a robot to the human."
So I think sometimes the term is a, as the industry's used in manufacturing. So every human getting their own robot was suddenly become 20, 30, 40, 50% more efficient. But therefore, robotics is deployed across a larger group of workers. And that's what we call, I think we're showing here, is the attended robot. The attended automation, so that there is a human attending or there's a robot attending with the human. And they're actually automating sets of tasks within an entire process.
So an example there where we had one organization that had 280 people doing about 20 different processes. We didn't try to automate everything. But the net effect is that we about halved the number of people needed to do that by just automating many of the things that all of them do, as opposed to trying to do all of the things that just a few of them do. And that's where the RPA market has really been pushing is that because the technology can only do the unintended. They can only do 100% and that's where the disappointment is coming.
Ying Chen: That's interesting. So what it sounds like is when it comes to RPA, even though the terminology tends to refer to those kind of rules-driven, no intervention required automation, what you're saying is that there's actually two different kinds of approaches that you can leverage the same technology for? Where one is completely unattended, meaning it doesn't require any intervention, versus this desktop version? Or am I hearing that the approaches are there, but it doesn't sound like all vendors can support that kind of a workflow.
Francis Carden: Absolutely not all vendors. In fact, not very many vendors can support the attended. If you like, the assisted, augmented worker that gets their own robot, because they are using much more of the traditional screen scraping approach. And it tends to be a lot slower. So if you try to automate a task that takes 30 seconds for a human to do and it still takes 30 seconds for the robot to do, then the humans just going to be sitting there, arms crossed weight and get very frustrated.
So the vendors have switched to this RPA, going after that low hanging rote work in the back office, or in the outsourcers, where they're actually saying "Okay, let's just automate the entire end-to-end process." And so, technically, they cannot do and move that technology because that was where it was built over to the attended piece.
And that's where our technology actually started, was actually started in that world, but of course, automating 100% if it's available, we can do that too. So it's a clear differentiator.
Ying Chen: Well, so you know, just want to push back a little bit on that because it seems like screen scraping tends to be the dominant technology of choice. So if that's the case, so when you talk about this performance difference and the ability to be able to support this more like an ad hoc like workflow that you're talking about, where it's not completely straight through like experience. So tell me about well, what's the alternative technology? And why would you see a performance difference between a screen scraped robot versus a alternative technology-driven robot?
Francis Carden: Well, keeping it at the highest level, and try and get it so that our audience can actually understand it, is that one approach requires no interaction. In order for the robot to do the work, it's like, leave me alone. If you dare touch my screen whilst I'm automating, I'm going to throw my toys at my and stop. And that's typically what screen scraping technology is. So therefore, it has to automate the entire end-to-end process. The opposite's true. If you've got someone on the phone to a customer in a call center, and they're going to do the change of address. And at the point they're changing the address, the robot can kick in and instantly take over 30 seconds of their work and do it in one or two seconds. That's a dramatic impact.
And so, from a technical perspective, the RPA vendors have been stuck in this much more myriad technology of just let me automate everything, which sounds good, right? We're going to talk about in a bit. Let's automate 100% of everything. That's the unattended. And if you do automate 100% of everything, gotta love that. But the problem is there isn't that much to automate 100%. That's where the disappointment comes in. Whereas if you can automate 10, 20, 30, 40% of everybody in your organization, the benefits are much more significant.
Ying Chen: Got it. So obviously we have these different approaches that we're talking about, do you kind of mind taking us through in a little bit more details, as you think about the technologies themselves, as well as the specific use cases?
Francis Carden: Absolutely. So if you think, you know, I started off saying RPA unattended, very successful. There are great use cases for it. There are also many use cases where people have started down the path and as you said, have been disappointed. Okay? But you need to look for that ideal work that's designed for RPA. Once you build an automation in unattended fashion, you put it on a machine, you lock it in the cupboard, and it runs all day long, until somebody changes that application and then it stops, and the work stops.
So whereas in the attended world, you've got a robot running on every desktop, if for some reason, the application changes and nobody knew about it, the robot might stop. But that's really just think about the automation stop. Well, the human's still there. They can just carry on processing it until maybe this afternoon or tomorrow morning or overnight, somebody could say "Oh, the application changed. I've got to rebuild my automation." And it still might only take 10-15 minutes to rebuild that automation. But it hasn't stopped the work during the day.
And if you're going to roll it out to 10,000 desktops, and every one of your 10,000 or 50,000 employees is gonna have a robot, you can't afford for it to stop. Whereas in the RPA world, it will stop, and then someone has to come back and fix it. And again, there's nothing wrong with that, if as long as doesn't matter there's a bit of queuing going on. That's all fine as well.
But I go back to that point about 100% of the work. Everybody's buying into the fact that I'm going to automate 100% of the work. But that impacts only a small group of workers within an organization. I actually knew one customer that went off, did a pilot and actually said "Well, we've saved two FTEs." I'm sitting there thinking well, actually, if you use the attended version, you potentially it could have impacted 5,000 FTEs by 20% and that's a significantly greater number.
Ying Chen: And is that because typically you don't think of an organization comprised of people who are doing only highly repetitive work that's high volume driven, that's all rules-based, right?
So to your point, maybe the two FTEs may be concentrated around a specific process, like fulfilling compliance requests, whereas some of these additional impact points that you talk about where you're saving maybe like two minutes, five minutes, as in Kerim's demonstration of Pega World. You know, logging in to your desktop every morning. And that's across thousands of workers. So is that the kind of differentiation that you're drawing?
Francis Carden: Yeah, and in the back office, you may have people doing reconciliation. You may have 200 people doing reconciliation, but there might be 100 different ways of doing reconciliation and that's why we have many people. And to try and automate every permutation of every type of reconciliation, you start with one and that's two FTEs have saved.
Whereas if you look across that entire group, so, we could automate how to get the data, how to get ready. Let them process those reconciliations in a robot-like manner, let the robot takeover, but the human sitting there, almost babysitting it a little bit. So you end up with a significant savings.
So we're going to talk about a use case in a minute. But if you go to the next slide, I think that really helps us actually describe what's going on. So if you look on the left hand side, we're looking about the number of people that's being impacted. So think about front office workers, or large groups of back office workers, that represents most businesses and most cost. And that's the bump in the middle there, right?
And if you look on the right hand side, that's the back office. Now, if you just move the graphic on, that really is the difference between attended and unattended. A lot of the analysts have picked up on this now, because they know some of the RPA vendors or many of the vendors simply cannot do the attended. So they're forcing you down "Let's go and automate 100% of the work." But we don't have as many workers in that back office doing that, that doesn't require a lot of work.
So nothing wrong with it. In fact, the technology that's out there, that's real value. But that's where most of the market has been with RPA. But what we've been doing for the last 10 years is actually both of those. So if you use the technology that can impact both sets of the organization, front office and back office, then you can get a significant benefit. And you can also do it much faster because in order to automate 100% of a process, you have to know what 100% of the process is.
But if you want to automate 20 things across 5,000 people, you could automate those literally within say, six to 10 weeks, deploy that out to everybody, and then continue on that level of improvement over time. You don't have to sort of wait until it's all done, you can actually go in that. Well, everybody talks about agile in a very loose fashion, but literally in an agile fashion roll out my automation as I've done them. And so, I'm impacting my workers straightaway.
Ying Chen: So again, just want to push back a little bit because from a market perspective, again, all of the reports show us the RPA market, where again they're focused on this unattended set of operations using robotic automation. It's gaining quite a bit of momentum and market share. So I guess can you actually do a side by side comparison to show us what the differences between adopting a RP-only approach versus a blended approach?
Francis Carden: Absolutely. So yeah, so I've prepared a couple of slides here, just to talk about it. So the first, on the left hand side, so it's a telco, okay? Both very large telcos. On the left hand side, a telco with 25 million customers or more, and on the other side, even smaller, 15 million.
But if you look at the impact over five years, they say they have a couple of hundred robots running, saving a few million dollars. That's a use case that's public and in the market. But we had a customer that actually within the first year deployed robotics to 20,000 workers. So it's a massive difference. So impacting 20-30% of 20,000 people by using robotics RPA, but attended, and that's been running live for the...well, actually over 10 years and saves that customer over 10 million a year.
And all the time, a new application might be on-boarded or end of lifed, and they can modify those automations to keep up with that change that's going on in that telco. So you can see it's not that the left hand side is wrong. It's the fact that if you could think about doing both of them, you actually can have a greater impact in a shorter amount of time.
And the second one is a bank. So first one was a telco. And I know we have a lot of people on the phone today coming from the banking industry and financial services. So this is a very large bank that's been doing robotics for actually nearly seven years. Is a very large bank, over 100,000 employees. And in that time, they told us they were only using unattended RPA. And in that time, they only went live with just about 200 robots. You think about that. And the average is one robot is what? I don't know, one to three FTEs. So you can work out the numbers.
But in another example where we went to one bank, very large bank as well, but we deployed in less than nine months, 300 robots, basically one to every worker, and they need half the number of people in that group to do the same amount of work. And in nine months versus what? I'm saying seven years to get a significantly greater impact.
Where there is ripe work for RPA unattended, do it. But try to find something that can allow you to do both. Don't just park yourself down one avenue, which is to do only unattended.
Ying Chen: So the numbers are quite astounding. I mean, you have this one case where this particular bank spent what? Five years trying to get live with 200 robots. And then on the Pega side, you're saying 300 robots within nine months. The numbers are not adding up. So help me understand how you can have one scenario where it took five years to do 200 robots versus nine months, 300 robots.
Francis Carden: So use the old adage. It's the 90/10 rule, or the 80/20 rule, whatever you want to call it. To do a percentage, 100% of somebody's work, it's really hard to do. I go back to this point, if it was easy, it would have been done already. So someone's work is just hard to automate.
But if you look behind anybody's shoulder in your organization, you will find at least 10 or 20 things that if you just automate that, you could get 20 to 50 to 60% saving. So we focus on that just automate that. Automate the easy road work that goes on in within a complex process and let everybody in your organization benefit from RPA that way with the attended. And if you get to 100%, then you take that work and you put it in the server room. You can do both.
Ying Chen: So basically, the difference here is the use cases, right? So to your point, it's really hard to find those complete 100% levels of automation, whereas I think that this case that you're talking about where they're leveraging a combination unattended and attended, so now the 300 robots that we're really looking at are all those different use cases that can be applied towards. Is that..
Francis Carden: So it's 300 robots running on everybody's desk, but actually running many processes that are repeated across all 300. They don't all do something. They all login, they all look for stuff, they all do balances. So there's lots of stuff that's repeated across that group. So we focus on that first, and that's why it was delivered in nine months. That was 20 different processes. But just automating and I think the lowest was 30%, is what we got with it. Doesn't sound very high, but that was across a group of 70 people. And the highest we got was 80%.
Now they're going back and actually saying "Okay, out of those 300, how many of them can now be RPA unattended?" And some of them can be. Some can't. Some require a signature verification, so that might be left in a future for AI, but for now, it's a human that does that. You know? And also some address changes, you know, still requires that human to oversight it.
So it's very important to realize for RPA unattended, you have to do it all bar none. You could break the process up and maybe do chunks of it, but then you're changing the way people work and that get very difficult for enterprises to do. It's the balance of automating some of the work across a lot of people as opposed to trying to automate everything across a few people.
Ying Chen: Actually, I think it's really interesting because if I hear you correctly, it's almost like the robots in some of these combination scenarios are acting more like humans. They're doing multiple types of tasks. They're not doing one step in the process. They're not waiting for our batch of workers to complete their tasks before they kick off their own work. Am I hearing you correctly?
Francis Carden: No, you're absolutely right. That's a good way to put it. I mean, primarily, we started lighting the contact center. And in the contact center, it's so variable, isn't it? You don't know what somebody's going to ask you. You typically are working on seven, eight, 10 different applications at the same time. You spend forever looking for things. It's a computer. Why should you be manually looking for stuff. Right? Let the computer bring that forward very, very quickly.
And so, that's why we see a 20 to 30% improvement in the front office and a 20 to 80% improvement in the back office. You know, as you go towards that back office, the work tends to take longer, so your automation benefits are greater, but impacting a smaller number of people, right? You still might impact 100 people, and if you can get 80% of 100 people in nine months, as opposed to 100% of maybe 10 of those 100 people in five years.
I mean, that's the difference really. It's what I want. I mean, it's how you go about doing integration. You don't try to boil the ocean overnight, and RPA is the same. Don't try to boil the ocean. Make sure you can do the unattended and the attended in parallel to get the benefits from both.
Ying Chen: Yeah, let's not boil the ocean. I never understood why you want to boil the ocean in the first place. So-
Francis Carden: You never get there.
Ying Chen: Exactly. So by the way, so let's talk a little bit about how our customers or prospects can think a little bit about how to actually get to that state that you're talking about, where they're leveraging both types of scenarios.
Francis Carden: Yeah, so this is an interesting one because if you look at all the processes, I think it's a great diagram here is that, you know, everything looks like it can be automated now. Like robotics is here. It's like you look around and you're hearing from the analysts-
Ying Chen: Artificial intelligence, you know, 60% of the jobs could essentially be automated, 40% of the task of all jobs could essentially be automated. And yeah, I mean, it sounds good, right?
Francis Carden: Well, it doesn't give any credibility to the fact that IT and business have been trying to do a lot of this for many years. So it's obviously not going to be like just suddenly come along and do that. I mean, we'll have nobody working at all in the next five years if you listen to some of the RPA vendors and what's going on in the market.
So really, it's important to have this checklist, right? About where you're trying to get to. You know, everybody says "Well, this is not about FTE replacement." I don't mind whether it is or not, it's about automating and stop paying people to do things that robotic automation can do for you. So I think, you know, define and focus what it is you're trying to achieve, in what groups are trying to achieve.
Don't just look at this as a one way street, only RPA, because many of you will be disappointed that in six to 12 months, you'll go "Well, that was good. But we got 20 FTEs saved." I mean, you want something that's going to have more applicability across a larger part of the business.
And understand the differences between this attended and unattended. The RPA vendors that do not do attended will absolutely, and rightly so for them, steer you down an unattended only route. If they cannot do the attended automation, they have to steer you down you must do it this way and this way only. And whereas, when we get people back on track, we normally get "Wow, we didn't realize you could do that." And then they get that greater success in much shorter time.
And I think it's just also go beyond the small scale thinking. We get people come to us "Oh, we're just trying three pilots. And we're going to spend six months doing those pilots." Well, in six months. I can actually improve large groups of people by significant percentages and have that in production and live in very short order.
And I think that's really one of the key things that you need to understand. Don't be fooled by any of the RBA vendors, us as well, right? I mean, we've got over 200,000 robots running live in production, running billions of transactions every single day. And we've got customers that will talk about it.
But everybody's got great use cases. It's really a matter of saying which use cases is the most applicable to your business, to what the outcome you're trying to achieve.
Ying Chen: Which brings me to my next point. So it's like, first of all, the technology's been around for 30 plus years, but suddenly it feels like everybody is in the RPA game. I think you can get our RPA from-
Francis Carden: What is it 30 or 40 vendors suddenly?
Ying Chen: Well, even from your BPO vendors, right? I know that BPO use robotic automation. There are folks who are specifically within the contact center that's leveraging robotic automation, and there's obviously pure RPA players and again, you know, technology seems to be very similar for the most part. Screen scraping, right? So how should our customers make sense of all of this?
Francis Carden: Well, let's be honest. Nobody wanted to do this, right? I mean, over the last 10 or 20 years, we've been promised solar. So we've been promised all sorts of transformation. And the great companies are on that journey. All they're looking for is a way to get some benefit in the short term. Whilst that journey maybe is taking longer than they expected, or maybe why they're just starting on that journey.
We put up a list of questions here, right? These are the things you should ask your vendor. Really challenge them because a lot of the RPA vendors have got some great use cases, but when you dig deep, they're great over five years. Or they're great, but they've end of life. So really dig deep into the case studies. You know you don't want to see a 20 to 30 robot that have been taken a year or two to deploy when you know that you've got other opportunities as well.
So robotic automation on the attended side improves dramatically customer experience because you're on the phone for less time, or you're spending more time on the quality outcomes with the customer. So you have more opportunity to upsell. It's very hard to upsell someone if you're on the phone with them for 10 minutes. Certainly, because it taking so long to sort their problem out, right? So that's a very key point of it.
And actually have people demonstrate the ability to actually automate whilst a human is still working. And also then compared to speed because the speed's going to be very important in an attended manner. Less important in the unattended manner because who cares if a robot takes the same length of time as a human if I'm not having to pay the human. But it starts to become important next year and the year after, right? 'Cause otherwise you'll have too many of these robots.
And also look to a vendor, say "What next?" You know, these robots are tactical and anybody that says robotic process automation is not tactical, I don't think deserves any credibility at all. I've always been selling robotic automation for the last 20 years if we can use the rename, but it is tactical. You know, I'm proud of the fact we sell tactical. It solves as a bridging technology. But the companies that aren't looking to transform as well out of that are going to end up with still a massive cost five years from now, keeping the lights on, running these legacy systems. You know, that's a huge expense in BIU. You know, basically keeping these applications and these legacy systems maintained.
So robotics doesn't solve that problem. But transformation does. So think about the journey three years, five years out about what you want. You do not want 500 robots running five years from now, just replicating the same work, sometimes often the same bad processes in the same way, but just having a robot do it.
Ying Chen: But something that you said that's really interesting because we get this question a lot, right? So when you say tactical, right? It feels like there is a near term timeline associated with tactical robots, right? But the same time you mentioned earlier about these automations that happen in place at our customer sites for years and years. So help me understand. So is it tactical and near term? Or is it long term and strategic?
Francis Carden: Well, I think tactical, agile, and strategic.
Ying Chen: Okay.
Francis Carden: But since we've been acquired, as you know, I was one of the founders of OpenSpan, the robotic process automation technology that started life in the attended world and into the unattended world, we've been doing this for a long time and our customers are using it. I'm not standing still. They've gone through major upgrades of their CRM systems or their back office systems. And the robotic automations can be very quickly modified and adapt, and some of them whilst they're still running are end of lifeing some of those applications as they go.
One day, maybe they'll have no need for robotic automation in the back office for some of those processes. And it is a law of diminishing returns as well, which is why I'm excited about being part of Pega because we can take customers, not just on that tactical journey that I've been selling for the last 20 years, but now a tactically, strategic journey. And I think it's a very important differentiator between us and our competitors.
Ying Chen: You bring up a good point because I think looking at some of the analyst recommendations when it comes to RPA, I see a lot of advice out there in terms of not confusing the automation and the work that you're doing around RPA with true business transformation.
And I think that that's what you're speaking to where we have heard scenarios where customers have implemented with another vendor robotic automations that take 24 minutes to run. And you're kind of wondering what's taking so long, and it's basically because they've never taken the time to do this transformation that you're talking about, where they actually consolidated their applications, maybe leverage sort of our case management mentality about getting to the outcome.
And so yeah, you may have automated the work, but your customers are still essentially waiting.
Francis Carden: Right.
Ying Chen: Right? So, and that's really-
Francis Carden: It is a key differentiator. I'm not saying, you know, it's a bit like do it now because you need the benefit. You need the saving. There's a lot of driver on BPO, we have the picture of the three little piggy banks there, right? Because people really do want to cut costs. But those applications that don't get re-engineered over the next five years, the competition isn't coming from the bank next door. It's coming from the bank that's not even started yet. Two years from now, that bank will come up and won't carry on that legacy baggage that you're wrapping with a robot.
So wrap with the robot, get the benefit, but start using some of that funding to drive the transformation journey you need to be on. And I think it's tactically strategic is an argument you're going to hear me say till the death, but it is truly tactically strategic that you need to be thinking.
Ying Chen: Gotcha. So let's see, we've got some more advice here.
Francis Carden: Yeah, just the last few bullets. Okay. So just make sure that, you know, figure out whether you are trying to replace the FTEs or not. Are you trying to improve customer experience? And you can have both. I think that's an important part. A lot of the RPA vendors say it's not about cutting FTEs. Well actually, that's fine. I still go back to that point where but if you can save a lot of FTEs, you can decide whether you want to move those FTEs to go do other work elsewhere. Or you can take on more work if you're a growing enterprise.
But I'm seeing a lot of the vendors saying "Oh yeah, we haven't saved any FTEs". Well, you don't really see that with our robotics. You know, there are a lot of customers who save a lot of FTEs as well. You know, where I'm deployed out to 20,000 desktops, and we're making 20,000 people 20% more efficient, do the math. Are they still found more work for 4,000 people? Probably not. So I think that's an interesting one to look at it.
And also do look at the advantages of RDA. And when we float that to their freely, but go out and actually look. It is very different. Your RPA vendors don't want to even talk about it. So we used to call it RDA, robotic desktop automation. I think that the industry analysts are now calling it attended and unattended, which is why you may hear me say the terms interchangeably.
And then, as I said, that transformation journey is very important. Think about where you're going three to five years from now, because IT will say "Hey, listen, if we do robotics, what happens to our transformation journey?" It cannot stop. You know, you will not be an enterprise that can compete in the market that we've got changing more rapidly than anything else I've seen today unless you think about robotics and the transformation journey as well.
Ying Chen: Great. Well, thank you so much, Francis. This has been an amazing conversation.
Francis Carden: It's always fun.
Ying Chen: We have received a lot of questions during our webinar and we thank you for them. But we want to be cognizant of your time today. So we'll be getting back to you within the next couple of weeks.
Well, thank you so much, Francis. And thank you everybody for your time and we look forward to getting in touch with you.
Francis Carden: Thanks Ying. Thanks, everybody.