Robotic Process Automation (RPA) tool adoption has soared during the past couple of years. If anything, the pandemic gave organizations added motivation to automate in general – especially around mundane tasks, the type that are in RPA’s sweet spot.

Gartner recently reported that spending on RPA software will top $1.5 billion this year and continue to grow by double-digit percentage terms for the foreseeable future. Deloitte has previously projected that RPA will hit nearly universal adoption in the business world at some point in 2023. There are lots of other lofty stats and anecdotal evidence that tell a similar story.

Don’t expect that story to change in the new year. Gartner, for example, expects spending on RPA software to grow nearly 20 percent in 2021.

[ How can automation free up more staff time for innovation? Get the free eBook: Managing IT with Automation. ] 

Key RPA trends for 2021

Amidst the continued fever pitch around RPA, we asked a range of experts to share their insights on what the new year will bring. Here are five RPA trends they think are likely to unfold in 2021. How do they fit into your RPA and automation plans?

1. Adoption shifts to measurement and (re)evaluation

Alex Day, SVP of Americas at Signavio, thinks that the outsized assessments of RPA adoption are spot on. Growing use is all but a given, and it’s essentially yesterday’s news.

“The question in 2021 is not whether RPA is going to grow or not,” Day says. “Instead, what we will be evaluating is how effective these RPA initiatives have been.”

That’s because the focus to date for many organizations has simply been getting some form of RPA up and running, finding what the technology is capable of, and learning some early lessons along the way. It’s one thing to get a bot into production; the barrier to entry there is actually relatively low. It’s another to determine whether your RPA strategy will actually deliver meaningful results.

“As is the case with many emerging technologies, there’s a possibility that the RPA ‘bubble’ will burst if many deployments do not go as smoothly as planned,” Day says. “As a result, what we will see next year is a re-evaluation of whether many of these RPA projects were architected effectively or not, and whether the underlying business processes were thoroughly understood enough to automate effectively at scale.”

[ Related read: How to identify Robotic Process Automation (RPA) opportunities. ]

2. Much heavier focus on process

Robotic Process Automation is something of a mixed-up term. For starters, there are no robots involved. (Disappointing, but true.)

And because automation is such a white-hot topic in IT and business circles, it tends to get an overweight amount of attention in the context of RPA. The word process is just as important, however. RPA on its own will do nothing to improve a process; it will just enable it to run automatically.

[ Get no-nonsense definitions of RPA: How to explain Robotic Process Automation (RPA) in plain English. ]

Expect a sharper focus on understanding and optimizing processes as a direct result of the shift from RPA adoption to evaluation and optimization. Plenty of organizations will realize their initial efforts were stymied by processes that they didn’t fully understand or that simply weren’t good fits for RPA.

Day predicts that process-focused technologies and practices – such as process mining – will gain a greater share of attention in the new year. Related terms and technologies such as process discovery, process intelligence, process optimization, and process orchestration will similarly become a bigger part of the RPA vocabulary and toolkit. And as we wrote about recently, we could see a closer relationship between business process management (BPM) and RPA going forward.

[ Read also: How to explain Business Process Management (BPM) in plain English. ]

“Most companies are jumping straight into RPA or trying to automate processes without first adopting process mining, which leads to more strategic deployment of RPA and a more efficient automation framework overall,” Day says. “By more closely associating RPA with process mining and process management, RPA will stand a better chance for success – and organizations will not adopt automation for automation’s sake, and instead focus on ROI and higher success rates.”

3. RPA use cases increasingly expand beyond finance and accounting

Finance departments quickly became fertile terrain for RPA’s growth in the enterprise. They’re full of the types of repetitive, computer-based tasks (such as invoice processing) that are often good candidates for automation.

Finance certainly isn’t the only home for RPA in business contexts. Jon Knisley, principal, automation and process excellence at FortressIQ, expects RPA and underlying process improvements to spread faster into other business functions as earlier adopters look to extend its use cases.

He points to customer experience (which may go by different names in different organizations) as a chief example.

“Customer experience is an emerging area for process intelligence [and RPA] as companies look to extend beyond the traditional finance and accounting practice areas,” Knisley says. “Companies today employ hundreds of applications to ensure high levels of customer satisfaction and it is critical to be able to look across many applications to identify potential bottlenecks and sources of friction.”

Looping back to #1 and #2 above, automating a bad customer experience will exacerbate the problem. Companies that start to apply RPA to different business functions will be best served by making sure they first understand what they’re automating.

“Before you can improve the experience, you have to understand the current state of user interactions with the company,” Knisley points out.

HR is another traditional department where RPA stands to expand its applications, especially because it still commonly involves a lot of manual, repetitive effort.

“The overall hire-to-retire (H2R) process integrates a mix of disparate systems that often require swivel-chair work to migrate data across systems throughout the employee journey,” Knisley says. “Process intelligence can optimize the process by identifying opportunities for workflow automation and re-engineering.”

In general, expect organizations to build momentum from early successes by replicating them in other areas of the business. Moreover, expect companies to try to move into higher-order use cases, aided by growing integrations between RPA and other technologies (more on that in #5 below.)

“The first step in digital transformation for many organizations was to deploy RPA for tedious and time-consuming tasks such as data entry,” says Chris Huff, chief strategy officer at Kofax. “In 2021, organizations will level up and increasingly leverage the power of RPA for higher-value use cases to transform their business workflows.”

4. Data privacy becomes a bigger concern for RPA

The extension of RPA use cases to departments like HR highlights another likely development, according to Knisley: Growing concerns around data privacy and security as they relate to RPA.

This isn’t just a function of expanding usage, but also of the likely increase in how organizations deploy related technologies such as process mining, or what Knisley refers to as “process intelligence.” As it gets easier to discover data across systems and automate the movement of that data with RPA and other technologies, the risks grow.

“Traditionally, data mining applications have been limited to core back-office ERP applications that generated log files to serve as the prime data source,” Knisley says. “With the emergence of process intelligence and data capture at the user interface layer, the variety and volume of data have increased significantly. The richer data set enables more meaningful insights to be uncovered, but at the same time the potential for data leakage increases as well.”

This will become particularly important as RPA is increasingly linked with other technologies that enhance its core capabilities, both at the process layer (see #2) and the automation layer. For the latter, think cognitive technologies such as machine learning. Knisley points to massive advances in computer vision as another example. Data privacy and security will be critical.

“Companies have an ethical and moral responsibility to address the use cases that can abuse those same advances in technology,” Knisley says.

5. RPA gets tighter with complementary technologies

Both the process and automation layers of RPA will become more integrated with other technologies as end-users and RPA vendors look to build on the basic capabilities of RPA software. This will be key to automating those higher-order workflows that Huff from Kofax noted above.

The general theme here is greater integration among processes and systems, as well as greater integration and alignment between complementary technologies, such as process mining, RPA, and AI disciplines such as computer vision or machine learning.

“As these next couple of years unfold, this stack will take on new levels of integration and capability,” says Stephen DeWitt, chief strategy officer at Automation Anywhere. “As enterprises re-cast themselves as digital natives, processes based on outcomes and contextual automation for their customers and workers will be fundamental by the middle of the decade.”

The phrase “intelligent automation” has sometimes been used to describe this shift; it reflects that RPA on its own is not a cognitive technology, but there is potential value in making it “smarter.” As a simple example, an RPA bot tasked with copying data from one source and pasting it into a specific field on a web form will typically break if there’s a UI change to the target field unless someone also updates the rules the bot follows to reflect that change.

DeWitt and others see far greater possibilities in the future based on more integration between complementary technologies. DeWitt even anticipates the beginnings of “autonomous automation” in 2021 – that is, bots that can themselves automate processes. Today, IT teams often need to intervene to connect to various apps, interfaces, and databases to retrieve information or apply rules to complete processes, Dewitt says.

Plenty of experts note that AI and automation done right will still require a lot of human interaction and oversight. But the relatively mundane RPA bot of today may soon be able to do much more of the heavy lifting involved in process automation itself.

“By the middle of the decade, or in some cases, next year, we’ll be able to automate the process of automation,” DeWitt says, “and once this moment happens, the enterprise will be able to rewrite itself.”

[ Is your culture agile enough to handle RPA? Get the digital transformation eBook: Teaching an elephant to dance. ]