From the robots in Amazon warehouses to the Internet of Things, automation is becoming more commonplace in sectors previously unoccupied by robots. The financial benefits of automation are predicted to be nearly $2 trillion globally.
Financial services firms have embraced automation on the front end, implementing solutions such as algorithmic trading. However, the potential for cost savings, better service, and legal compliance extends throughout the organization as well, with firms looking to automate repetitive, business, rule-driven work. But how?
Enter robotic process automation, or RPA—a rapidly growing solution for back office automation poised to reach $1.2 billion by 2021. Amidst this growth and hype, it remains important for firms to take a sober, yet open-minded look at what RPA can do for them. To this end, Celerity is beginning a series of articles on RPA and how it might benefit your organization. We will focus on what RPA is—distinguishing it from other related processes—and analyze methods of RPA commonly found in the real world.
What RPA is and isn’t
CIO.com defines RPA as “an application of technology, governed by business logic and structured inputs, aimed at automating business processes.” Mindfields Global defines it as “a software robot or application/tool that can be configured to perform tasks normally performed by a human, using rule-based processes.” PwC calls RPA “small, lightweight, easy-to-program software tools that can automate a range of digital activity.”
So think of RPA like a low-code Excel macro, but for your whole computer. An RPA robot can log into Citrix, access a remote desktop, enter an internal company database, copy the records from that day, paste them into Excel, sum the revenue, make a chart, and email this report to the head of the finance department. And unlike a human employee, it can do this all day, every day, without sick time, without vacation days, and without mistake or oversight.
RPA is nested within the broader category of intelligent process automation (IPA), which “describes a range of automated processing, from RPA all the way to cognitive technology and machine learning systems” (PwC).
Since RPA simply involves robots doing what humans do on a computer, it must be distinguished from Straight Through Processing (STP), “where there is no visibility to human eyes of the steps/action taken by an automated process.” Because RPA robots simply live inside the computer they’re working on, they should be distinguished from “walking, talking auto-bots” or “physically existing machines processing paper” (Deloitte).
Importantly, RPA is different from business process management, or BPM. IBM notes that BPM is “an approach to process operations management which focuses on improving process performance by streamlining business processes, removing bottlenecks and adding value.” One can easily imagine using RPA as a way of fulfilling the process-streamlining goals of BPM. Think of BPM as a toolkit, and RPA as a tool in the toolkit.
RPA is also different from artificial intelligence (AI)—which Google defines as “the theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.” RPA is a program that automates routine tasks according to fixed rules, whereas AI is a capability for machines to engage in tasks that require human-like thought. If RPA can follow a recipe you give it, AI can look at what’s in your pantry and offer dinner suggestions. A future article in this series will discuss how RPA and AI can be combined and are being combined.
A taxonomy of RPA
To shed more light on RPA, we provide categories for classifying different flavors of RPA using real-world examples. The first way to classify RPA is whether a human is still involved in the process (attended automation) or not (unattended automation).
Unattended automation occurs when an RPA robot can, without prompting by a human, initiate and complete a business process. Unattended automation is a purer form of RPA, since it requires no human interaction, but it should be emphasized that this is only possible if the process being automated is sufficiently simple.
Attended automation means human intervention is still required at some points in the business process, but the RPA robot automates several steps. Employees can still perform tasks too complex for the robot, which is left to perform the manual, repetitive tasks. While this is a less pure form of RPA, it still brings benefits in cost saving and frees up humans to focus on the thought-based tasks at which they excel.
RPA can also be defined by the type of automation that is occurring. While this list is by no means exhaustive, it may serve as a starting point for company leaders considering where to start looking for processes to automate.
The processing of clearinghouse payments is sufficiently routine that it can be automated. A Fortune 100 diversified bank recently used RPA to automate its ACH (automated clearing house) payment processing.
RPA can be used to automate certain routine customer requests. A leading European food producer used RPA to automate the processing of simple customer order inquiries, reducing manual effort by 40-60%.
RPA is also useful in performing tasks involving validation, as it is easy for computers to tell if one thing is not exactly the same as another. As an example, BNY Mellon automated web-based client record reconciliation, freeing its employees from a tedious, mistake-prone process.
Robots using RPA are also able to assist in pattern detection, detecting anomalies or, at the very least, packaging and delivering information to human workers who can do so. An RPA vendor client automated segments of fraud detection by having the robot check databases and present a report to a fraud prevention analyst, more than halving the time spent on requests.
This article has opened the door on RPA—we know what it is, what it’s not (for now), and where it fits into the automation space. We’ve seen several ways to classify RPA, all the while providing real-world examples of the use of RPA to drive tangible, dramatic benefits. In the next post, “Why Use RPA,” we’ll discuss the benefits of this technology.