By automating simple, low-variation and high-frequency data entry, RPA can produce impressive results for companies. Judging the true value of RPA has been difficult, however, as RPA adoption trends have produced high expectations.
Companies need to understand the limitations of RPA technologies before making significant investments. The value of RPA can still be great if its deployment is deliberate, focused and utilized as one piece of a larger puzzle.
Behind the rise of RPA adoption
RPA has its roots as on-premises software. It's generally implemented within legacy systems. At its core, RPA is a programmatic macro that emulates clicks and keystrokes on a desktop computer, according to Sean Chou, CEO and co-founder of Catalytic, a Chicago-based process automation software company.
When it comes to simple and repeatable tasks, RPA can make business practices more efficient and allow a human employee to focus on more complex tasks. One of the most common RPA adoption trends is using it to transfer information from Excel sheets and other documents.
"In many ways, RPA has been successful because so many corporations have been slow or unsuccessful at modernizing their systems," Chou said.
The continued reliance on legacy systems at many enterprises has created an opportunity for RPA tools. Deploying RPA on legacy systems can be less expensive than going ahead with a full digital transformation.
"Businesses always talk about modernization and going digital, but the reality is most enterprises aren't quickly replacing or modernizing most of their legacy systems," said Burley Kawasaki, chief product officer at K2, a process automation software company based in Bellevue, Wash. "And there's a lot of remaining manual input into older applications."
It is in this way that RPA acts as a type of bridge across the technology gap between legacy systems and modern digital applications. It can complete business processes rapidly and accurately at a lower cost than an employee completing the same work.
Companies will look to get the most out of RPA
Still, the limitations of RPA are very real.
"Long-running and variable processes, as well as processes that involve multiple human interactions and handoffs, are challenging for RPA," Chou said. "Bots are designed and optimized for a short, continuous activity stream with little to no variation."
When there is too much variation and the running cycle is too long, the odds of a break happening increases. Overwhelming the bot like this leaves the company back at square one.
"RPA lends itself to a few use cases very well and can often deliver some eye-popping ROI with an initial, cherry-picked set of use cases," Chou said. "But between the limited use cases, the costs of development, the infrastructure costs, and the much higher than anticipated maintenance costs, business cases have been challenging."
While it holds potential, RPA must be utilized more as a piece of an overall strategy.
"To some extent, I believe RPA will always be somewhat limited when operating entirely on its own, without any human interaction whatsoever," said Harel Tayeb, CEO of Kryon Systems, an RPA and AI vendor based in New York. "At least for now, the technology requires an element of human interaction to deliver the maximum benefits."
Living up to the hype
To overcome the limitations of RPA, enterprises need to be strategic about how they implement the technology.
"As you think about automating ... think about all the tools you need for intelligent process automation," Kawasaki said. "RPA is a very good tool, but it's just one of probably a number of things that you need to really look at."
While RPA can make legacy systems more efficient, companies should be careful not to let this dissuade them from updating and evolving their technology infrastructure.
"To address these shortcomings of RPA, organizations should take a moment to rethink their long-term automation strategy, especially if their strategy consists entirely of RPA," Chou said. "Automating processes should start with the process, and not a task in front of a computer. It should connect and amplify people, not emulate them."
But as hype around the technology builds, it has become a victim of its own success. Companies need to keep in mind the limitations of RPA and find the best implementations, rather than looking for it to solve every problem.
"The RPA hype has been fierce, and I think it'd be impossible for anything to live up to the level of growth and funding we've witnessed in the past four to five years," Chou said.