When you hear the word “robot,” you might envision a near-sentient being from a science fiction movie or perhaps the lovable household helper Rosey from the Jetsons cartoons. While we’re a ways away from humanoid robots or helpful office assistants, the idea of “intelligent” tools that can help us complete useful tasks is arriving in the workplace.
Various tools that fall under the general category of robotic process automation (RPA) are emerging and worth investigating. These tools are not physical robots but digital technology that mimic mouse movements and keystrokes while “reading” the data that appears on a screen. They can make basic, rules-based decisions and alter how they perform based on those rules.
A simple automation might “read” the fields in a column on a spreadsheet of overdue invoices, create an email with a copy of the invoice attached, and send it to a delinquent customer. More sophisticated automations might receive incoming emails and, based on the text, either update some data in your accounting software, route the email to another department with additional information, or flag the email for human processing.
Many of these tools are designed, at least theoretically, for non-programmers, and several have free versions available or are included with cloud-based software like Microsoft’s Power AutomateTM.
In a job environment where talent is hard to find and retain, RPA can automate routine and uninteresting tasks and free up expensive humans to do more valuable work. RPA also never sleeps and can be used to perform work after hours and often at higher throughput than humans.
Getting Started With RPA
Broadly speaking, there are two approaches to getting started with RPA, which may be performed simultaneously if desired.
IT Led Approach
The first is to centralize RPA deployment: IT selects a standard RPA software package, then designs, deploys, and maintains automations for other parts of the company.
This approach is likely comfortable for most technology departments since it follows the same flow as typical application development or enhancement projects. Requirements are gathered from potential users, then tools are developed, tested, and ultimately deployed. These tasks can be performed internally or with the help of a trusted partner. An IT-led RPA implementation is generally appropriate for more complex automations, as the benefits of using your tech talent to solve a problem are tempered with the longer timelines required at most IT organizations.
User Led Approach
The other approach is to provide tools and some basic training to users that might benefit from RPA. Some of the user communities that can benefit most from RPA are those that deal with multiple systems and spend a significant amount of time doing basic data manipulations between systems. You can often identify these groups as departments full of people using dual monitors with a “data providing” application on one screen and a “data receiving” application on the other.
While not everyone will be willing or interested to learn RPA, you’ll likely find one or two individuals who will embrace the technology once they see the benefit for their job. Target one of the simpler tools that uses a graphical or web-based interface or provides simplified training to get started.
Help these individuals identify a simple automation task to start, and assign them a “buddy” in IT to help answer questions. You might be surprised at the excitement generated when a non-tech individual develops a simple automation that might do little more than copy data from one spreadsheet to another. However, this can be the start of significant time savings with minimal time investment from IT.
The Challenges of RPA
There are two primary challenges to RPA in its current state. The first is that, like many newer technologies, the gap between marketing and reality can sometimes be wide. Vendors imply that “anyone” can quickly train their automations, and that they offer human-like intelligence.
The truth is that most of these tools are still rules-based and require you to define explicit rules that inform their activities. An RPA cannot determine whether a line in a spreadsheet represents a long-term customer that should be treated carefully or a new customer that repeatedly fails to pay their invoice unless there is a clearly defined rule to make this distinction. Technologists and business users accustomed to complex spreadsheets should be able to understand this nuance. Still, those who aren’t used to distilling requirements into logical rulesets may expect more than the technology can deliver.
Training and development tools for RPA can also vary. Most tools include some sort of graphical training component but ultimately require the user to delve into some form of programming for all but the most rudimentary automations. This might range from flowchart-like visual programming to using a common programming language like JavaScript.
The second challenge is deciphering which applications your RPA tool is compatible with, especially as several vendors offer multiple versions of their RPA tools. Microsoft’s Power Automate, for example, has a web-based version that’s fantastic for manipulating files and components of the Microsoft cloud ecosystem but cannot open an application on a user’s desktop and lacks integrations to some other cloud services. The desktop version of Power Automate can open desktop applications and manipulate applications, but has a different interface and slightly different capabilities.
It’s worthwhile to experiment with tools and technologies that you may already have as part of other subscriptions. However, if you’re going to invest heavily in RPA, make sure you understand the nuances of the platforms you’re considering and whether they’ll work with the applications and tools your users will need to automate.
While we’re unlikely to see cubicles filled with smiling automatons anytime soon, RPA is a valuable tool that’s worth investigating. It’s easy to start with low-cost automations using a tool you may already own, explore complex IT-led automations, and empower your users to build their own automations for routine tasks.
