Micro Focus is now part of OpenText. Learn more >

You are here

You are here

Robotic process automation: 5 key trends to watch

Ericka Chickowski Freelance writer

Robotic process automation (RPA) technology is on fire, and RPA adoption by enterprises shows no signs of slowing down. Analysts say that in the next 12 months, RPA growth will outpace the overall tech market. But most companies that have adopted RPA are already experiencing the kinds of pains that you'd expect from a technology still sitting at the peak of its hype cycle.

RPA is undergoing a redefinition of its place in the enterprise—and of the product category itself. And many organizations will be busy this year re-evaluating the tools and processes they're trying to automate and brushing up on the kind of governance and organizational coordination necessary to scale up and mature their RPA efforts.

Here are five key RPA trends to watch this year.

Organizations will struggle to scale RPA

RPA adoption is expanding rapidly. The software will hit $2.4 billion in sales in 2020, according to Forrester Research, a threefold increase from 2017, and Gartner named RPA the fastest-growing technology market last year.

But that rapid expansion in RPA adoption was primarily for pilot projects and isolated use cases.

While many organizations have raced to try out RPA in small doses, few have scaled up the automation technology across their businesses.

The issue of scale is the No. 1 thing that preoccupies people about RPA right now, said Forrester analyst Craig LeClair.

"Some 50% of companies—and these are companies that have some maturity in RPA, not the newbies—have fewer than 10 bots," he said in a recent presentation at the Institute for Robotic Process Automation & Artificial Intelligence. And while some organizations have shown that it's possible to scale up with 500 bots or more, they're the exception.

To date, fewer than 2% of businesses have scaled to that level, LeClair said. Most are still struggling to figure out how to put more bots in place across their organizations. And many organizations are still puzzling out how to increase the utilization rate of their existing bots—how to get them working more frequently to improve the ROI from those purchases.

Forrester found that only 29% of organizations are satisfied with RPA utilization rates, with 62% of respondents reporting that their bots are working their prescribed tasks fewer than three hours per day.

RPA orchestration will move to the forefront

Several headwinds are keeping organizations from scaling with RPA, according to LeClair. Many of those come back to the tricky challenge of managing and orchestrating bots.

"The problem with RPA is that you're creating these new process flows and creating a need for infrastructure to automate all that," said Archie Roboostoff, functional and performance portfolio director for application delivery management at Micro Focus.

The more bots that organizations build to automate tasks, the more they bump into time-consuming complexities that paradoxically arrive from trying to automate away worker tasks in the first place. If not done well, Roboostoff said, it is possible to spend more time managing the bots than they will save.

"You've essentially created a new monster with RPA that you didn't have before, which might have negative consequences to your original intent."
Archie Roboostoff

"You might end up saying, 'Wait, we went with RPA to clean everything up and make it more productive.' Well you did that, but as a consequence, you now have to manage 600,000 different process flows that are all fragile at times and now you've got to hire a whole team to do that."

An RPA solution needs to be resilient enough to withstand the changes to the applications you're automating, Roboostoff added. "Otherwise, you’ll have a flow management problem keeping pace with that."

This issue is at the heart of what Hfs founder and CEO Phil Fersht calls RPA's "void of promise unfulfilled," and it's why it has become clear that the market urgently needs redefinition and a new manifesto.

"The discussion noticeably switched to end-to-end process and workflow optimization rather than the whole 'bot for every desktop' nonsense," said Fersht of 2019's sea change in RPA philosophy. "2020 will see a new narrative emerge as smart enterprises explore a broader tool box of process discovery, process mining, process automation, and data ingestion to help them orchestrate true end-to-end processes that actually bring an enterprise’s customers, employees, and suppliers closer together."

Process discovery will remain top-of-mind

Chief among those management and orchestration activities in 2020 will be the push toward greater maturity in process discovery—essentially in vetting processes for potential automation in RPA, mapping out the tasks related to the processes, figuring out how they tick, and documenting the flow.

"You're starting to see more emphasis on the overall process discovery, getting a little bit more fundamental on the front end," Roboostoff said.

Turning process discovery into a science will be crucial not only to ensure that RPA bots designed to automate processes are robust and resilient, but also to weed out less applicable use cases and ensure that organizations are building RPA automations that maximize utilization rates.

"Many companies struggle with figuring out what they should automate," said Aniket Maindarkar, vice president and general manager of the business process services practice at DXC Technology. A big part of his firm's RPA services are designed around an initial discovery process that examines historic transaction data—be it accounts payable or accounts receivable data, HR onboarding transaction data, or something else—and uses AI tools such as Google TensorFlow to make sense of how frequently transactions occur, how long steps are occurring, and so on to identify the most applicable RPA use cases.

Enterprises will broaden their automation lens

Digging deep into the process discovery journey will reveal uncomfortable truths for many organizations. A big one is that often the time-consuming processes they think are ripe for RPA are taking too much time because they're bad or poorly conceived processes.

As organizations mature in their automation approaches, they're increasingly taking a big-picture view of how to solve business problems, rather than simply identifying tasks or processes to target with RPA bots.

"What is most challenging for buyers is identifying what business problem to solve, and that goes down into, 'Am I solving a task problem? Am I solving a process problem? Am I solving an efficiency problem? Am I solving an integration problem? What problem am I solving?'" said Dave Easter, director of product management for RPA at Micro Focus.

"A lot of times when you solve those task problems, you're uncovering a lot of other much more systemic problems with the process itself."
Dave Easter

Easter described one recent experience where his firm was helping a bank that wanted to design RPA automation around a process that had people go into a financial database, page through data, find disputed transactions, copy them into Excel, and send an email message to a manager. The bank simply wanted to automate those tasks, but his team had it take a moment and re-examine the validity of the tasks themselves. Easter's team helped the bank recognize that, for compliance and security reasons, it shouldn't be routing the information through Excel and email anyway—it needed to make some important process design changes.

"That's the kind of larger vision that you need to be able to bring ... beyond just that initial task automation," said Easter.

Some analysts, including LeClair, call this broadening of scope "intelligent process automation" (IPA), which includes the process discovery analytics and orchestration components mentioned above. But Easter wonders if IPA is a bit of a buzzword that covers functionality crossovers and integrations in existing RPA and business process management product maps. Either way, it's clear that 2020 will bring a much broader look at RPA's role in the overall automation landscape, rather than simply deploying in isolation.

Governance will grow in importance

The broadening vision and a drive to scale RPA in the enterprise will result in organizations taking a more mature approach to RPA and automation governance in the coming year.

"Because RPA has been coming in though the business, there's been a lack of control and governance over the use of more standard tools for shared services across the organization," said LeClair. Governance, he said, is more of an organizational issue than a weakness or opportunity in any given RPA platform.

He sees more companies taking a "strike team" approach to automation orchestration as a way to amp up governance capabilities while still providing flexibility and alignment with the business. These teams will include stakeholders from across the business who will formalize pipelines where the business can come up with automation ideas and run those through the strike team, which will then run process discovery and review those ideas.

"They're creating almost this crowdsourced approach of looking at opportunities applying to automation, and then putting in place guardrails and guidelines that can ensure that ideas have been validated, vetted, are viable projects, and are well-designed," he said.

This is a maturation year for RPA—and process automation

Ultimately, the trends unfolding in the next year presage an evolution not just toward a maturing of RPA but toward broader automation across the enterprise. Your organization will need to bring governance and orchestration to the table to keep all of its automation plates spinning, and you'll need to scrutinize how your manual processes work today if you expect to purpose fit automation not just to your current processes but to the underlying business needs. That's ultimately the key to evolving RPA from initial pilot projects into full-scale deployments.

Keep learning

Read more articles about: Enterprise ITIT Ops