question marks on trees
Robotic process automation is in its early days and therefore causes confusion. Here are some answers to your FAQ automation questions PHOTO: Evan Dennis

Forrester Research predicts the robotics process automation (RPA) market will reach $2.9 billion by 2021 from a base of $250 million in 2016. That is a substantial leap for a technology that, in some quarters, is still poorly understood. Or to put it another way: RPA is an emerging technology whose full potential still goes unrealized by many people.

RPA, which is also called robotic digital workplace, can easily be confused by the uninitiated with traditional automation, a decades-old technology. To clear up any confusion between the two and any other misunderstandings associated with RPA, we have answered some common questions about the respective technologies.

What Is the Difference Between RPA and Traditional Automation?

Traditional automation is the automation of any type of repetitive task. It is usually found in a product workflows but can be as simple as a process for screen scraping. It entails application integration at a database or infrastructure level and can take months to implement. 

RPA is another form of automation, and has several distinguishing features, Parikshit Kalra, SVP, Solutions and Capabilities at business and service provider HGS told CMSWire.

  • It is non-intrusive. Most often, the RPA bots stay in the front-end of the system and moves forward with the task, without having to intrude on the back-end.
  • It is system agnostic, working across application types.
  • It is able to take action quickly due to its ability to mimic the role of an agent.
  • It is easily scalable and system integration is easy.
  • It is easy for non-technical people to use — there is no code to learn; it is almost all graphics.

Perhaps most significant of all, RPA is very quickly implemented, compared with traditional automation, which can take several months. Also, unlike traditional automation, RPA doesn’t require application integration. Instead, said Sasanka Panda, business unit head of Business Process Solutions at IT solutions organization NIIT Technologies, RPA uses the graphical user interface, or GUI, to perform its tasks across multiple systems.

When Is it Better to Use RPA Versus Traditional Automation?

RPA is a quick fix solution when compared to traditional automation, Panda told CMSWire. It can be implemented in a matter of weeks, while traditional automation takes months.

Sometimes RPA can be used in the short term, until a traditional automation project can be planned and implemented, he said. And sometimes, RPA is the best solution no matter what the time frame. For example, RPA is better suited for more personalized engagements — or any task, for that matter, that is more complicated to execute and requires access to multiple applications, Kalra said. RPA is also better for scenario-based tasks, he added.

Does RPA Require a High Level of Integration?

No. One of RPA’s distinguishing features is its ability to connect systems that cannot be easily automated through traditional automation approaches, Mark Snyder, senior manager of Operations Excellence at business consultants West Monroe Partners told CMSWire.

For example, one of West Monroe Partners’ RPA bots can pull new-hire information from a word document, and in turn use that information to populate ADP payroll, expense and a time-entry system, he said. “It then designs and orders business cards and even schedules building access on an employee’s first day. Traditional automation methods just can’t handle this type of scenario.”

RPA also comes in handy when automating legacy applications that don’t provide their own APIs, said Mike Fitzmaurice, VP of Workflow Technology at workflow automation software provider Nintex. “I’m grateful for RPA for situations where a vendor has failed to provide an API, but I feel let down when it gets to the point where I need it,” he told CMSWire. “Modern integration is API-based. All well-designed software should expose an API and a UI tier should be consuming that API.”

Will RPA Make Traditional Automation Obsolete?

No, said Kalra. He used this analogy: you don’t need an excavator if a shovel is good enough for the task at hand. “There are still numerous tasks that traditional automation is perfect for — simple calculations, linking one system to another — tasks that do not require advanced programming, handling of multiple systems and with a low cost of integration,” he said.

Traditional automation’s ability to move vast quantities of data, quickly, between systems remains unmatched — certainly not by RPA, Snyder added. “RPA only works at the speed of a user interface. It takes our RPA tool five minutes to load a new employee’s data at West Monroe, which seems dreadfully slow by traditional automation measures. However, when you compare this to the 30-minute manual effort it replaced, the value is tremendous.”

There is, however, the distant possibility that RPA could become a threat to traditional automation should overzealous vendors and/or CIOs decide they’d rather integrate their apps and content at the user interface level and abandon the movement toward standards-based APIs, Fitzmaurice said.

“It would be a tragic development after all the work that has been done on REST, OData, JSON, XML, OAuth, OpenAPI, etc. It would also put the responsibility for integration squarely back in the hands of a few vendors” instead of where it is right now: that is, with a professional developer.

What Comes Next for Automation?

Automation is moving along a continuum, Anurag Bhatia, SVP, global head of Business Process Services at IT services company Mphasis told CMSWire. Macros, scripts and workflows made up traditional automation. RPA is the second generation. And the next level, he said, is cognitive — or the ability to handle unstructured data.

For all its bells and whistles, the current iteration of RPA is missing cognitive AI enhancement. It can perform certain tasks, but it cannot make sophisticated decisions. Which is supposed to be the point, at least when the conversation turns to how many jobs automation and RPA will wind up eating.

“As technologies evolve, more and more of the mundane aspects of the job get addressed, leaving the individual to focus on high value, judgment and analysis focused activities,” Bhatia said.