Micro Focus is now part of OpenText. Learn more >

You are here

You are here

Robotic process automation: 4 trends your IT shop should watch

Esther Shein Freelance writer

In this time of uncertainty, many organizations have turned to robotic process automation (RPA) to meet changing customer needs, and the technology is expected to ride a tidal wave of growth for years to come.

More industries will continue to implement it, and in the next few years, RPA will incorporate new technologies that will serve to broaden its appeal to enterprise customers with opportunities for more advanced use cases. 

The global RPA industry is projected to reach about $19.5 billion by 2027, up from $1.6 billion in 2019, according to a report by Allied Market Research. While software and financial services organizations have been the earliest adopters, the services, healthcare, and pharmaceutical segments could see significant growth in the next few years, the report said.

Adoption of RPA technology is key to helping organizations to improve business processes and capabilities, create efficiencies, and reduce operational costs, the report said. On the other hand, growth could be inhibited by "data insecurity risk and less potential for RPA in knowledge-based" businesses, according to the report. 

Other experts agree that RPA growth is in the cards. A February 2020 Forrester study on the future of work and automation found that 48% of 160 respondents planned to increase their RPA spend in the next year by 5% or more.

That resonates with IT leaders, who are anxious to automate more processes. "We're on mission to relentlessly automate everything that's manual," said Jacqui Guichelaar, group CIO of Cisco. 

Here are four trends to watch as you implement RPA in your own shop.

1. RPA will incorporate process intelligence and cognitive automation

In the last few years, people have recognized the limitations of RPA—including its ability to scale and problems with orchestration. Vendors have started exploring the concept of cognitive automation, to leverage machine learning and artificial intelligence to make decisions where an RPA bot cannot, said Geoff Hodgson, director of the automation practice at global IT consultancy Saggezza. 

The industry has reached a stage where "almost" any process can be automated, Hodgson said. The question now is how to achieve scale and automation processes where the potential ROI is less than ideal, he added.

Process intelligence technology can help with this task. It mines a company's processes and tasks, and bundles in AI to build a data-driven picture of an operation, Hodgson said. That way, organizations can understand the pain points and potential bottlenecks while exploring an automation opportunity, "and how to process something in a safe digital environment before taking [a project] live," he said.

David Casas Alarcón, accounting lead at CBRE Property Management in Warsaw, Poland, said his firm is using optical character recognition (OCR) for supplier invoice management and uses built-in RPA functionality to complement this.

When vendors send their invoices, "we apply OCR to read and abstract the invoice information that is uploaded to the accounting system," Casas Alarcón said.

"Uploading information to the system can fail for many reasons, but a robot is trained to detect system errors, read the cause, fix the error, and re-upload the data." 
—David Casas Alarcón

The robot continues to repeat the process if the upload fails again until this is successful, he says. "In addition, the robot runs reports to ensure that all invoices have been successfully processed and provides daily process status."

As organizations progress in their adoption of RPA, they tend to become more ambitious with cognitive technology, which mimics human functions, according to a report by Deloitte. More than a quarter (28%) of those implementing and scaling RPA are also implementing cognitive automation, while only 6% of those that have not implemented RPA are progressing with cognitive automation, the report said.

2. RPA won't fully replace all jobs 

The concern that jobs will be seized by robots has bothered humans since the Industrial Age, said CBRE's Casas Alarcón. But, he said, people should view RPA as an opportunity to shed themselves of mundane tasks better suited for robots, he said.

The digitization of finance processes really brings "a transformation possibility for roles that have remained unchanged for decades," he said. "Far from being a threat to finance employees, RPA assists [with] opportunities of more well-paid jobs," while automating "soul-crushing corporate tasks and helping users to concentrate on being extra creative and perform thought-provoking work."

Not everyone agrees that all knowledge workers are safe from the RPA bots. A recent story in the Nhe New York Times cited a McKinsey statistic that says displacement is on the rise. The consultancy had predicted, pre-pandemic, that 37 million US workers would be dislocated by automation by 2030; McKinsey recently increased that prediction to 45 million.

A lot can change in a year. Most respondents to Deloitte’s 2020 Global Human Capital Trends survey said they viewed AI mainly as an automation tool and a substitute for manual labor rather than as a way to augment or collaborate with humans. However, this year's survey found that executive respondents believe the use of technology and people is not an "either-or" choice but a "both-and" partnership.

"The application of automation will inevitably lead to changes in the way businesses structure themselves. So someone doing role X may end up losing role X—but people will not be replaced entirely by automation."
Geoff Hodgson

How bots and humans can get along

Hodgson believes the objective of automation will be to make humans as efficient as possible rather than do away with their jobs altogether. There are always going to be exceptions to that, he added, but "I'm quite hopeful that those people who lose jobs to automation will be able to find better work than keying in data from documents."

Further, Hodgson maintains that call center jobs that were shifted offshore will come back to the US as a trend toward "relationship management" takes hold over the long term.

"The human touch, as it relates to client and customer relationships, will be key differentiators. The value of having that meaningful human interaction will be higher than it is now."
—Geoff Hodgson

While acknowledging this may be a more expensive proposition than fully automating call centers, Hodgson insisted that it behooves organizations to do this. "You want to feel like you're having a conversation with your neighbor, with someone who can relate to you locally."

3. The rise of no-code/low-code/citizen development

Citizen development, largely driven by low-code and no-code technologies, is becoming a big trend, said Hodgson. And this, in turn, is "pushing companies to start to explore how they can put automation in the hands of the knowledge worker."

As part of implementing citizen development, organizations must ensure they have the necessary governance and that "what you're allowing these people to automate is not too complex and far-reaching," he said.

However, a level of automation can also be applied to work that is somewhat complex "if it means you enable people to make better and more consistent decisions across the organization," Hodgson said.

For example, automation can be applied to the job of an underwriter, "a human embodiment of corporate risk appetite," he said. Automation can help them make more consistent and better underwriting decisions. "In that instance, you're not automating boring tasks but something to help people be better at their jobs."  

Nate Tsang, founder and CEO of Hong Kong-based WallStreetZen, agreed, noting that teams can get more done without having to develop in-depth RPA technical expertise to respond to any given issue.

"RPA is currently best utilized as a tool for citizen developers in low-code or no-code environments. The goal is to bridge the gap between these citizen developers and the complexity of their tasks using RPA."
—Nate Tsang

Organizations are past the point of looking for easy-to-automate tasks such as data transformation, Tsang said. "There are already many options for things like that. Rather, RPA needs to become more fully integrated into the development environment to the point where it can be used by nontechnical staff."

4. RPA will become commoditized

Enterprises that are not yet at least experimenting with RPA are "at risk of significantly falling behind the curve," Hodgson said.

It might get easier, or at least less expensive, to catch up. Obtaining a software license from an RPA vendor could be "prohibitively expensive" for small companies in particular. This year, Hodgson expects enterprise vendors including Microsoft, Google, and Amazon Web Services to begin offering less expensive RPA services.

These services will boast "capabilities that are enough for small and medium businesses to start seriously exploiting automation technology," he added.

Automating tasks doesn't change based on the size of a company; the only difference is smaller companies are doing the same amount of work but with fewer people, he said.

The road ahead

Hodgson predicts a fascinating battle between legacy RPA and BPM (business process management) vendors and the industry giants, all coalescing around a core set of capabilities and vying for the same share.

Within the next five years or so, there will be one or two large companies left standing after everyone else is swallowed up by the tech giants and the large RPA vendors, Hodgson said.

"They’re all marching toward the same common set of capabilities. [Niche vendors] will find it hard to survive in the long term."
—Geoff Hodgson

Keep learning

Read more articles about: Enterprise ITIT Ops