"Govern Your Bots!" by Loreal Jiles (Strategic Finance, January 1, 2020): https://sfmagazine.com/post-entry/january-2020-govern-your-bots/

Links to learn RPA Development or other things about RPA Tools for Free 

FULL EPISODE TRANSCRIPT
Adam: (00:00)

Welcome back to Count Me In, IMA's podcast about all things affecting the accounting and finance world. This is your host, Adam Larson, and I'm happy to bring you episode number 40 of our series. Today's conversation is between my cohost Mitch and Loreal Jiles IMA's director of research for digital technology and finance transformation.  Their discussion revolves around RPA and the practicality of robots and accounting. So, let's transition over and hear what Loreal has to say. 

 

Mitch: (00:31)

We keep hearing so much about RPA, but I'd like to kind of start off on the right page here. Can you tell us exactly what is RPA? 

 

Loreal: (00:45)

Certainly. So RPA stands for robotic process automation, and my personal definition is RPA is a technology that enables a virtual robot or a bot, as we call them, more casually to emulate human interaction with computer applications to execute processes. And so in planar terms, what that really means is the bot mimics the clicks and the keystrokes of humans to perform a process on a computer. It can navigate through desktop or web based applications, and it can do all of that simply by accessing what's called the user interface or the UI. And the user interface is the part of an application that all of us as humans see and interact with. There is another part of each application that's in the back end and that's where traditional computer science majors or programmers access that to right code to either modify the application or to automate things that way. Right? Historically, the only way to automate things was to access the backend of the application. What RPA does is it allows us to do things, things that historically businesses, professionals or non it professionals would not be able to do because now without having to access the back end of the application where there's admittedly a bit more risk, we simply record the keystrokes but clicks the type, of the human throughout the process and allowed the robot to replicate those steps. After we've documented it in sequential format in the required, I'd say the required format that's necessary. So what's really cool with it, I think is as opposed to going in the back end again, the automation happens right in the front of the screen. And in some instances you can even watch the robot perform your process. Um, so RPA for me is the easiest way to automate processes for non it professionals. 

 

Mitch: (02:42)

So you just said it right there. non-IT in our audiences, finance and accounting. So how real is RPA in finance and accounting? What are some of the typical uses that you've come across? 

 

Loreal: (02:52)

Oh, I think it's incredibly real. I'd say first off, it's, it's very real across many. And so McKinsey published a study just a few months ago highlighting RPA offers potential ROI of 30 to 200%, and in fairness with the right selection principles for a process, those projections aren't incredibly conservative. I start there because nearly all enterprises who have begun RPA journeys regardless of their industry, they started in the finance and accounting department. And so that means that the thousands of companies that are using RPA today, over 70% of them started in finance and accounting. And that's incredibly powerful because it says to us as finance and accounting professionals, how do we want to be impacted by the technology? Do we want to be the recipients of it, if you will, or to be on the front end. And my encouragement to finance and accounting professionals is to consider being on the front end of it. The other thing I think I'll highlight is, although it's tempting to say, um, we're not developing bots or are developing robots themselves or actually developing processes that robots will perform. And the only thing we have to do to do that is effectively to tell the robot what steps to take. Yeah. Who knows those steps better than the people that actually perform them. Yeah. So my mantra is more can we have finance and accounting professionals developing these processes as opposed to solely it professionals. That doesn't mean everyone will be converted to a developer, but you could play a critical role in either process requirements gathering in, in exploring what's needed to be able to facilitate or progressed the implementation, if you will. And so I think in that space, we're seeing something where every company, or just about every company that starting RPA journeys is starting in finance and accounting. Our processes lend themselves a bit more. It's a automation because of the cyclical nature because some of them are routine just by nature of what we're doing for month end close or quarter end close for example. And so I think in the spirit of that we as a profession will see a ton of this, for those of us who haven't already been exposed to it, significantly, 

 

Mitch: (05:10)

Well, in my own personal research trying to learn about RPA and how finance really fits in, I came across democratization of RPA and quite honestly, I'm hoping you can give us a little better definition than what I read online. 

 

Loreal: (05:23)

Oh, certainly. Yeah. So democratization of RPA is a phrase that summarizes the concept of a bot for every employee and some people go so far as to see a bot for every human. I'm not trying to go that far, but the concept of a bot for every employee means that if I sit in any corporate setting, despite the company, just like I have a laptop and it's automatic, that when I start working for that company, they will give me a laptop. And that laptop will have Microsoft office suite when it, for example, there's an by some people or envision that I will also have a bot on my laptop as well and I'll have software that allows me to automate some of my processes on demand and in real time. And my robot that sits on my laptop will effectively become my personal assistant in that world. I think we've, we've got a little wait before we can get there. So would I ask myself if I truly believe in democratization of RPA? Yes. I think it's a far off, but in fairness, no one thought we would have a laptop when every desk, just a few short decades ago, no one thought we'd have effectively a computer in the palm of our hands or in our back pockets. And so what I think we have to do before we can truly realize something like democratization of RPA is that the technologies themselves, the software vendors, we'll have to continue to progress the user friendly nature of their tools. There are a couple of tools that are more user friendly than others. And there's a recent Forrester report that was published earlier this quarter that tells us who the leading RPA players are. And I think the top three that they've highlighted are UI path automation anywhere in blue prism. Having touched each of those systems personally, I've found the barrier to entry much lower for UI path for a non it professionals. And so that answer's very different if you're thinking about a world where it professionals will develop all of the opportunities themselves. But as we progress along this journey toward democratization, that means we would have to have less and less dependence or reliance on it, expertise or computer science and knowledge to be able to progress. So I think in this world, in today's world, I don't see it happening in the next couple of years. However, if these, these software vendors continue to progress along their journeys and they're moving incredibly fast, RPA is a relatively new technology, just a few years older. So yeah, if they continue to progress along their road-maps along the journeys, then yeah, it very well may be something we see in the next five to seven years. 

 

Mitch: (08:03)

Well, you just referenced three pretty specific, you know, software programs and very tech savvy businesses. So what is the reality for, you know, business in general? Uh, how can businesses kind of take advantage of these different RPA, processes and, and what are you seeing as far as some of the smaller companies trying to take advantage of this? 

 

Loreal: (08:26)

Sure. So I think that a couple of things for people to realize is what these vendors have done now is they've provided training for free. So for companies like UI path and Automation Anywhere there is online training available in their websites for free. And that means the same training that we take people through at my company today in order to get them certified in RPA development is the same training that someone sitting at home who's unemployed could take for free on the internet. And so I think that's the first piece to realize that the, the data and the information and the tools are really right are at your fingertips. I think the piece beyond that, as we think more strategically about an organization implementing it for larger enterprise, several of them are already starting when their journeys, they have whole it organizations, the it organizations are standing up the infrastructure dev test and prod environments. They're installing the software on virtual machines, right? They've got a small dev ops team that's working through the implementation and if they're doing this the right way, then that dev ops team is a hybrid of it and business professionals. So that's what we see most customary in larger enterprises for smaller than mid size companies. What I'm a bit startled by this, it doesn't seem that they've embraced the technology as much. And I'm not sure why that is. I don't know if it's because there's a belief that it costs too much or there's a belief that it's only for larger enterprises or that all this infrastructure is actually required when if you go to the website and download the application for free, the software for free, you'll get the same enterprise grade software that we have at our company. And so in that world, if you could put that on your laptop, learn how to use it for a smaller company, I think there's actually stronger business cases for small and mid companies to leverage this groups where there are 10 to 20 or even 30 people that are in their finance and accounting department. They're generally overworked because they're doing end-to-end stuff instead of what you see in the large world where we've got teams devoted, yeah. Only to AP P teams devoted only to to AR to accounts receivable, and so in a world where you've got smaller than mid size enterprise, they're working into end and they've got a ton that they need to have happen in a very short period of time. RPA can be a perfect solution for those enterprises. Yeah. Yep. Price point is incredibly low. So I'd say at best they, a smaller group could start with maybe 30 grand or less per year in licensing costs and they could get the training for free, get a bit of support to figure out how to get up to speed. But outside of that, you're paying nominal things in licensing cost for a bot that can run 24 hours a day, seven days a week, 365 days a year. And that means in a world where you've got a team of 24 people, well who are working insane hours during close. If you could take away some of those activities that they're not all that excited about anyway, then you can increase your accuracy, you can increase the assurance for those processes and certainly the efficiency of being able to progress it. So I think at the world for larger enterprise is starting to catch on already. Along the smaller and mid size world, I think we've still got some ways to go and just making sure we heightened the awareness to this and what its true capability is. 

 

Mitch: (11:56)

You've certainly done a great job, you know, answering my questions and giving me a better understanding of RPA and you've done a really good job highlighting all the benefits, but I'm sure there's also risks that come along with bots like this. So what kind of risks do you think our listeners need to be aware of and twofold, how can businesses who are looking to implement this RPA mitigate those risks? 

 

Loreal: (12:18)

Certainly, certainly. So there are tons of risks with any technological implementation and at things that external auditors worry about. For example, are things like credential management, how will we manage the, the user credentials of the robot itself? Will it be a service account? Well, we treat it like it's a human account. When it needs access to the accounting system, what types of things will we allow the robot to have access to? Do the segregation of duties still apply in that scenario? Just as it does with the human, will we allow the robot to enter banking credentials for a supplier and to process an invoice and approve it for payment? Other things would be kind of bigger than that outside the tech space specifically is humans orchestrating the inputs to the process for a bot to perform and the process that they perform ends up being a fraudulent transaction. So those types of things are very real. And I think there was the first documented case of fraud for RPA just within the last year or so. And it was not so much on the credential management side, but humans orchestrating the inputs so where they know what the inputs to the processes, the bot picks up those inputs and then performs the transaction accurately, but with fraudulent information. And so those are, are some of the key risks that I think, especially as a finance and accounting professional myself, that I think about and that external auditors are thinking about on a regular basis to mitigate those risks. At BP specifically, we've written what we call a governance and operational framework. And that truly speaks to not just how will we govern which processes we will automate, but from a technology perspective, how will we govern, the credential management, we store all of our credentials in an encrypted vault, which some software vendors make available. And then there are others that you could patch on to that, that kind of reinforced that security, if you will. And there are a host of reviews that are required before we go live with any process. So one thing to bear in mind is from an implementation perspective, you could have centralized, decentralized or federated, models for your RPA program. And what that means is centralized would be a world where there's a small team or a large team in fairness that handles all development and implementation. If you go all the way to the other extreme on the spectrum, you could see a world where there's federated, where you've got a central hub that's maintaining governance, but there are smaller groups within the organization that are deployed and doing the implementation as well. And then all the way to the other end would be the world of citizen development where you truly put the bot on each laptop and empower the employees to build things as they desire and deem appropriate. And so different governance models would be required for a centralized versus the federated versus the true citizen development world. And I think that so truly mitigate that. It starts very early before you start deploying lots of opportunities. It starts with engagement with your program sponsor, which is likely a finance leader, right? That's relatively senior or at the executive level. It starts with key conversations with the it organizations in the digital security groups. We also proactively engaged our external auditors just to understand what risks they saw available or that might come up. And then figuring out how we mitigate those as well. And so I think the ultimate response to how we mitigate the risk is through governance of RPA. Right. And that governance should be tailored to whether or not there's a preference for a centralized model, decentralized, or the citizen developers option. 

 

Mitch: (16:06)

Well, there's certainly a lot to consider as we're talking about RPA and I know just all other digital technologies as well. We're progressing so fast, it's really hard to predict what's next. But I'm just wondering if you had any thoughts on, you know, putting these governance principles and policies in place, you know, how will that ultimately help a business in their implementation as RPA or any other technology, like I said, continues to evolve and how can organizations prepare for what the future may hold in the accounting and finance industry? 

 

Loreal: (16:40)

Sure. So I think that no digital technology will be truly implemented successfully and sustainably without governance. So you would do what we call a proof of concept initially to figure out if the tech, the specific digital tool, whether it's RPA or otherwise, if it works for you, if it integrates well with your applications that are within your organization, web-based or desktop applications. So you'll test out what we call a proof of proof of concept. And that's where you'll pick a use case and just see if the technology even works as a trial run. What's next I think is we generally do a bit of a proof of value. And so that's a concept where we say, let's make sure that this is actually going to be valuable. Will we save enough hours to make an impact for our teams? Will we truly be able to automate the processes that really matter or is whatever the technology is best suited for a different use case that's not quite as impactful for the organization. The step beyond that, and so what we did a bit in parallel proof of value is standing up the governance and so you'll want to find out that specific organization what the current ways of working are. If you need to do something like I did, which is write a hundred page policy for governance, or if you don't necessarily have to go that far, if you've got a smaller group where you have a 10 to 20 or so finance professionals within the organization, the the large policy might not be needed, but in its purest form. Every company that's getting started on an RPA journey needs to identify how will they handle technology governance? How will they manage credentials? They'll have to think through what their operating model will look like, how will they do process selection and identification and all of that. Regardless of if you're using RPA or data visualization or data science or data analytics as if you're allowing people to write Python scripts for analytics. If you go to other extremes where it's just your it organization doing artificial intelligence or machine learning implementations. And so with any of that, you'll require some governance. You'll require governance to identify what the right use cases are. Are you truly elevating the value right, that your finance and accounting organization delivers to the, to your broader business, right? Or are you focusing just on finance and accounting processes? And so what we try to do is strike the right balance there because ultimately we're in service of the broader organization. And so I think as we think digital technologies in the future, key technologies that finance and accounting professionals need to be aware of are data visualization, data science or analytics, RPA because it's got a pretty low barrier to entry, similar to data visualization and then yeah, an awareness level of surface level. I think also an artificial intelligence and machine learning capabilities, the impact that that'll truly have on our profession. I think we're still a few years out, but what we'll start seeing is I believe that data visualization and RPA, we'll have the largest impact in the next one to three years. What we will ultimately see, I think down the line is the artificial intelligence world where we're able to automate judgment. And so the way that we peer that with RPA is through something called intelligent RPA or intelligent bots. And so yeah, that's a world where RPA is leveraging other technologies, algorithms that someone may have written and the robot is then going to kick off running that algorithm to be able to then bring that data back and leverage it through some other automated process. It will be bacon, some of the judgment that we have now, so there's this concept or this thought that, Oh, right now we're only automating things that don't require any judgment. The technology is actually much farther along than people are with regard to implementing it and so there are already solid use cases for automating things that also require judgment. It's just integrating a more into end solution with multiple well technologies. I think down the line, what a finance and accounting professionals job looks like will not be, I had to do data entry, I had to get to work in the morning and wait for a report to refresh before I was able to perform any analysis. What we'll see in the future is we show up to work the morning of close and our robot has already run all of the reports we need to review. It's leveraged some artificial intelligence based on past trends to say, these are the exceptions that I believe needs to be reviewed, and when we get there in the morning, we'll be reviewing exceptions and spending more quality time on the analysis and how that translates into additional value for the business. 

 

Announcer: (21:30)

This has been Count Me In, IMA's podcast providing you with the latest perspectives of thought leaders from the accounting and finance profession. If you liked what you heard and you'd like to be counted in for more relevant accounting and finance education, visit IMA's website at www.imanet.org.