Micro Focus is now part of OpenText. Learn more >

You are here

You are here

Why robotic process automation has a bright future with enterprise IT

Ericka Chickowski Freelance writer

In the midst of unbridled enthusiasm for artificial intelligence (AI) and all things automated, enterprises are diving head first into the robotic process automation (RPA) market. Large organizations reap benefits of RPA through faster processing of data and less time spent by employees on repetitive tasks.

But many experts are also warning would-be buyers to keep in mind that RPA's upside isn’t limitless. They say it's a mistake to conflate RPA with AI or even simple machine learning (ML), and many believe that in a lot of instances RPA is a stopgap measure that might serve only until organizations can address underlying problems with business processes and legacy IT infrastructure.

Nevertheless, RPA still has great potential for enterprises in 2019 and beyond.

Evan Campbell, managing director of the technology consulting practice at Protiviti, said that as more firms move to digital transformation, it's a matter of progression. "You take the path starting with RPA, then move into leveraging more advanced forms of automation," which then gets you into ML, natural-language processing, and AI. 

"RPA is really the first jumping-off step toward intelligent automation."
Evan Campbell

Here's what your team needs to understand about robotic process automation.

What's behind RPA's sudden growth spurt

While RPA is hardly new technology, it's still a maturing market. According to Gartner, RPA spending shot up by 57% last year, hitting the $680 million mark for 2018. By 2022, analysts believe the market will reach $2.4 billion annually.

The quick-and-easy nature of RPA's automation of manual processes is buoying its growth at the moment, according to Gartner analyst Cathy Tornbohm.

While varying RPA technologies excel at different use cases and industry applications, the fundamental function of RPA is to automate defined, multi-step manual tasks that are done in high volume. In many cases these tasks stretch across multiple legacy technology applications or platforms that can't be easily linked.

Good candidates for RPA are processes that take structured data and transfer it across multiple systems, said Arup Das, CEO of Alphaserve Technologies. Other potential targets are time-consuming, low-volume activities that require frequent "system hopping"—for example, HR onboarding, 

RPA-enhanced data entry can reduce process cycles by anywhere between three and ten times, according to Vinot Tharakan of ClaySys Technologies. Take, for instance, the popular use of RPA in places such as IT help desks or customer call centers.

If you analyze most operational users or contact center workers, more than 50% of their day-to-day tasks are related to mundane data entry work, and if that were automated through RPA it could "significantly improve their productivity and the quality of their processes," Tharakan said.

The best use cases

The most beneficial use cases vary by industry, with a wide spectrum of creative RPA applications ranging from manufacturing to healthcare to finance. For example, manufacturers value the error reduction that RPA offers in back-office and operational processes, said Moshe Kranc, chief technology officer at Ness Digital Engineering.

RPA also helps manufacturing companies by reinforcing supply chain procedures and bridges the gap between redundant activities such as quoting, invoicing, accounts payable, accounts receivable, and general ledger operations. This adds "substantial value to the business outcomes," Kranc said.

Meanwhile, healthcare organizations are leaning on RPA to streamline clinical services and workflow for processes such as scheduling patients and insurance claim processing, Kranc explained. This delivers cost savings and an improved level of care.

Banks and financial organizations are similarly attracted to the cost savings and efficiencies offered by RPA.

RPA has become "an effective tool to maximize efficiency and keep costs as low as possible" while also maintaining maximum security levels in various processes, such as accounts payable and receivable, credit card and mortgage processing, fraud detection, underwriting support, service desk, and more, Kranc said.

Where RPA won't help

That said, RPA remains anything but a universal easy button for automation. There are plenty of use cases that experts warn enterprises to avoid.

First, organizations need to be sure they're not addressing a painful symptom with RPA, rather than a greater underlying business problem. For example, one of the big mistakes organizations make is to automate a fundamentally bad or obsolete process using RPA, said Henry Peter, chief technology officer at Ushur.

"While it is true that automation is at its best when certain repetitive tasks are automated with intelligent software agents, it also becomes the worst when that specific repetitive task in itself is outdated."
Henry Peter

You can use RPA to get immediate relief from inefficient processes, but organizations that stop with that level of automation are likely to only get 50% to 60% of the way there, said Alphaserve's Das.

"RPA needs to be coupled with process reengineering in order to create a permanent fix."
Arup Das

In the same vein, organizations might address the pain from a lack of integration between systems by using RPA to simplistically "glue" them together. But that wouldn't be a very strategic or long-term fix in this day and age of microservices and continuous delivery of software.

"We usually recommend staying away from processes that exhibit brittleness," said George Kaczmarskyi, who leads the robotics and intelligent automation practice for EY Americas Financial Services. Be cautious about using RPA to automate processes that undergo frequent business rule changes, that require interactions with too many applications or with applications that change frequently, that rely on data that has unpredictable quality, or that have small transaction volumes.

AI and RPA: Perfect together?

This is because, at their root, most RPA technologies are nowhere near as intelligent as ML- or AI-based automation. While some pundits and marketers use RPA and AI interchangeably, RPA technology usually isn't self-learning or self-adjusting.

"Most of the leading RPA vendors today do not have true AI capabilities. They are interfacing with other technologies that provide some of the learning capability—but most of the learning is happening outside of the RPA platforms."
—Evan Campbell

And the question that arises as more organizations mash up RPA with AI is: Is it even RPA anymore? Many experts say that this combination is what's spurring the rising field of intelligent automation.

"Intelligent automation is an evolved version of RPA, where AI aids and amplifies RPA technologies, allowing organizations to experience the best of both worlds."
—Arup Das

RPA can buy you time

Intelligent automation is still very much an emerging technology, and it will take some time for the technical stack to mature and the market to shake out. But if you decide to wait on that, you should still be working on technical debt and inefficient underlying business process—and there's still plenty of opportunity to derive value from RPA today.

At its core, RPA offers conservative organizations immediate automation options and frees them to wait for other tech to mature sufficiently, says Pankaj Chowdhry, founder and CEO of FortressIQ.

The ability to delay a major technology decision to see how the landscape evolves can mitigate costly mistakes, he said. Organizations should use the time afforded them by RPA to improve their technology decision making and optimize their architectures to use the best components available.

The fact is that there's never one true path for any technology at large, complex organizations. The journey to automation should be no different. Lauren Mahoney, director at Jabian Consulting, sees RPA as an important component of an overall continuum of process improvement and automation.

Mahoney said it is hard to say that any technology is permanent, but RPA can be an effective long-term fix in some use cases.

"In some use cases this technology may be enough, and in others RPA may pave the way to more sophisticated automation such as AI. However, you never know if the next innovation to solve the same business needs could be right around the corner."
Lauren Mahoney

Work your way up

Rather than dismissing RPA because of longer-term plans for modernization, organizations still invested in legacy systems that could immediately benefit from RPA just need to create a plan with realistic ROI projections, said Albert Rees, senior vice president and head of business consulting at EPAM Systems.

"The way we see organizations handling the 'temporary-stopgap' risk is by managing the RPA payback period."
Albert Rees

They should also consider adding some governance and scalability to the process by developing a center of excellence (CoE) that can centrally manage the decision making about where RPA makes sense to the business.

An effective CoE can provide the necessary governance, cross-functional alignment, and business case accountability that is critical to achieving the full value of RPA, Mahoney said. Once the CoE is established, it can be more proactive to identify additional potential use cases across the organization.

Read more articles about: Enterprise ITIT Ops