There seems to be an ongoing debate the past few years on whether or not Robotic Process Automation (RPA) is truly a useful tool for large scale use in businesses. One could argue that the undeniable saving on manual labor truly speak for themselves. On top of that, the possible positive changes in the overall quality and efficiency of process execution certainly can't hurt, either. Plus, the overarching thought of robots inherently means that you don't ever have to worry about the possibility of human error. So what, then, could the possible downsides of this wave of the future be?
The Primary Issue
Ernst & Young wrote in a recent report that, despite the optimistic outlook the company has on RPA and its potential for the business world, there are clear shortcomings. Coming from a company that has delivered RPA implementation projects to 20 countries and counting, this should mean something. They have cited that anywhere from 30-50 percent of RPA startup projects fail, with E&Y being redeployed to clean up the mess. E&Y writes that �this isn't a reflection of the technology.� The real problem lies with the companies involved.
Simplicity and the Problems That Arrive with it
RPA software was clearly designed with the intent of a simpler workflow in mind: a low-to-no code alternative to classic workflow automation. The process itself is easy: a robot examines a particular workflow process through the user interface (UI) and repeats the task as needed. Convenient, right? Well, digging a little deeper, it seems that the simplicity is something of a smokescreen. �A lot of times, companies will create a series of processes and consider it done. Really, the robots are constant time-consuming upkeep project.
This really seems to be one of the biggest repeating themes in terms of RPA implementation: the thought that �simple� means no effort. Because of the relaxed style of the programming, it's relatively easy for a less tech-savvy person to create a workflow process. Small picture, this is really nothing to be concerned about. However, looking at a larger scale, there is so much more that can go wrong. More people crafting workflows means anything going wrong can result it a massive setback just trying to reverse the whole process.
Changes to UI
One of the biggest ways that RPA can fail is the tiny breeze that can topple the whole building. Since the whole process is based off of robotic programming emulating what it viewed in the UI initially, any small change to the underlying UI will break the system. That's the other side of the double edged sword of robotics, really. A human could easily account for these changes and adapt, but a robot following orders can't exactly make the same spur of the moment judgement calls.
Even though it is a haphazard and cavalier way of going about things, there are often to automated regression tests set-up to fall back on in these moments. This is especially devastating when considering the massive quantities of bots in use for various automations. With all of these processes in place, it's extremely difficult to make up for any of the tiny issues that throw the whole system off. COO of Eggplant, Antony Edwards, has seen this before. �...they start automating, and either automate the wrong thing or get lost in trying to reverse engineer the process.�
RPA use in Legacy Systems
Another large problem that a lot of companies face is the debt built up by focus and upkeep on outdated legacy IT. Camunda CEO Jakob Freund doesn't speak very highly of current RPA status, believing it to hinder possible growth. �Applying RPA gives some executives the illusion they would actually be undergoing a 'digital transformation,' which would allow them to defend their market shares against new digital native challengers such as fintech startups or big techs.� The truth is that RPA just cannot make up for decade old IT infrastructure. Really, it could just be possible that without the painful lessons to help create necessary improvement, RPA in these scenarios can actually delay modernization.