RPA takes root as agencies use cases grow

Two years after NASA implemented the government’s first robotic process automation applications, a new report shows RPA has taken root at many agencies.

RPA program maturity grew 70% between fiscal years 2019 and 2020, according to “The State of Federal RPA,” a report published Nov. 1 by the federal RPA Community of Practice, which assessed 23 programs. The CoP has more than 1,000 members in 65-plus agencies. Its mission is to accelerate the federal government’s adoption of RPA, a low- to no-code commercial technology that can automate repetitive, rules-based tasks.

Additionally, the report found that the number of automations rose from 219 in fiscal 2019 to 460 in fiscal 2020, a 110% gain, and the annual hours automations save increased by 195%, to 848,336 from 285,651. That translates to 1,708 more hours of capacity per automation.

“RPA is more than just automation. It represents an opportunity to take a hard look at legacy business processes,” Gerard Badorrek, chief financial officer at the General Services Administration, said in an email to GCN. “Before we automate a process we apply lean process improvement methodologies to standardize and optimize that process. In some cases, we can eliminate a process altogether if the work and requirements have changed.”

Agencies use RPA for various business functions. The report found that the top three are finance at 49%, acquisitions at 14% and IT at 13%.

“Two key findings in the report suggest that increased adoption of RPA outside of the finance function will be a significant trend in 2021,” said Badorrek, who is also co-lead of Cross-Agency Priority Goal 6, which calls on agencies to shift from low-value to high-value work. “First, agencies are deploying agencywide center of excellence models that provide broad RPA services and lower the barriers to entry for individual programs operating within non-finance functions,” he said. “Second, more federal RPA programs are adopting enterprise-level technology solutions that enable a broader base of users.”

One driver of automation’s growth is the pandemic, said Jim Walker, federal CTO at RPA platform provider UiPath.

“Prior to the pandemic, they were looking at it as any other emerging technology,” Walker said, but then states got slammed with applications for unemployment insurance and food stamps, creating backlogs. As employees or their family members got sick or had to isolate, the smaller staff had even more work. “Automation really has shined as a way to help government permanently to solve its big challenges,” he said.

Still, implementing RPA has its challenges, the most significant of which is incorporating advanced and cognitive intelligence tools into solutions. Artificial intelligence, machine learning, optical character recognition and neuro-linguistic programming are the next step in RPA’s evolution, Badorrek wrote, but “many agencies are still at the starting line working to launch pilots.”

Another ongoing challenge is choosing RPA use cases and including process improvement and reengineering best practices into automation development, he added. The government is making progress: The average automation went from 1,300 hours of annualized workload in fiscal 2019 to 1,700 hours in fiscal 2020 – an increase Badorrek expects will grow in fiscal 2021 as programs take on more complex automations. 

A third challenge is securing RPAs, Walker said. A 2019 Office of Management and Budget memorandum, for example, requires credentialing of automated technologies involved in a transaction involving at least one federal system. This means that a bot needs credentials to send and receive encrypted emails, he said.

The report highlights five agencies’ success with RPA. At the Department of Agriculture, 95% of the RPA implementations are unattended and have realized more than 2,500 hours per bot. USDA’s RPA program has identified more than 70 future automation possibilities.

GSA is developing the first applications to be used governmentwide. They’re expected to be available in mid-fiscal 2021. Meanwhile, the program has created more than 200,000 hours of annualized capacity through more than 65 automations.

The Defense Logistics Agency has 96 automations in production that have saved more than 200,000 work hours.

Agency automations are “freeing up people from repetitive, mundane tasks that make up their workload,” Walker said. “Now they may have more time to spend on the phone with a veteran talking about their health care. They may have more time helping somebody navigate through the visa process or the passport process.”

The most encouraging long-term indicator of RPA’s use is the development pipelines the agencies referenced in the report, Badorrek wrote. On average, each RPA program has 10 to 20 automations waiting to be developed, and 10 programs have more than 30 automations in the pipeline. “This demonstrates the long-term sustainability of RPA, and the fact that RPA program customers understand the benefit it can have for their operations,” he said.