How RPA helps organizations deal with a global crisis

The coronavirus has caused priorities to shift. There is a new focus on automation tools that can be rapidly deployed to decrease costs, support the remote execution of business processes and build resilience.

Industry 4.0 / Industrial IoT / Smart Factory
Elenabs / Getty Images

One of the many consequences of the coronavirus pandemic is that it is leading organizations to find innovative ways to leverage technology — whether it’s to support work-at-home strategies, enhance customer service, or help combat the virus itself.

Robotic process automation (RPA) provides a good example. The technology is designed to speed up and improve the accuracy of repetitive tasks such as data entry, by using bots to perform these types of functions.

More organizations are deploying RPA platforms to help address challenges related to the pandemic, said Craig Le Clair, vice president and analyst at Forrester Research.

“The psychology and roadmap for automation has changed overnight,” Le Clair said. “Enterprises now need to prepare for a burst of investment activity that targets [cost reduction] and resilience but maintains necessary guardrails.”

Priorities have shifted to automation tools that can be rapidly deployed to decrease costs, support the remote execution of business processes, and build resilience. “RPA helps in all three areas,” Le Claire said.

Enterprises across many industries impacted by spikes in demand or special situations associated with COVID-19 that already know how to use RPA are adapting to support the excessive loads, said Maureen Fleming, program vice president and analyst at International Data Corp. (IDC).

The use cases for RPA during the pandemic are broad, Fleming said. They include network traffic monitoring, exchange of COVID-19 data, bank processing of government loans, managing the cancellation or purchase orders associated with cancelled orders, managing cancellations and refunds for air travel, and onboarding systems for remote workers.

“RPA is also used to speed up information collection and reporting to dashboards, customer portals and Web sites,” Fleming said.

“Whether a new form is created in response to the pandemic or an existing process, RPA is being used,” she said.

An example is registration to take a drive-through COVID-19 test. “In this case, the robot is able to perform the registration faster than a human, allowing healthcare providers to administer tests rather than fill out forms,” Fleming said. The technology can also be used by nurses for filling out infection reports that go to the Centers for Disease Control (CDC).

RPA can replace workers who perform certain tasks in front-or back-office positions, in many cases freeing them up to take on more interesting types of work. The automation capabilities can enable companies to complete projects much more quickly than they otherwise would.

For example, Le Claire said, one major airline received more than 120,000 cancellation requests in the early weeks of the pandemic, an increase from the typical 3,000 cancellations per month. A three- or four-minute manual review of each request was required.

The airline hired a company to build an RPA bot in just six days, with four all-night sessions, that could clear 4,000 refund requests per day. That’s more than would have been cleared in a month using manual processes, Le Claire said.

RPA is ideally suited to help organizations handle large jumps in data entry needs, such as in the case of having to quickly process healthcare data during a pandemic. “Data entry tasks are some of the easiest steps to record, analyze, and design a bot to replace,” Le Claire said.

County leverages automation in COVID-19 battle

Suffolk County in New York, one of the hardest hit areas in the country, is a case in point. The county in March 2020 deployed an RPA platform from UIPath with help from technology services provider SVAM.

The system includes licenses for up to three robots, and the county is using the technology to automate the processing of lab results, said

Scott Mastellon, commissioner at the Suffolk County Department of IT.

The county uses a New York State health system that provides it with lab information related to communicable diseases. This system provides each county in the state with COVID-19 lab results — both positive and negative.

Suffolk’s health department was retrieving lab results from the system, printing out the lab results, sorting through the results for positive cases, writing a sequence number on each positive result, batching up positive cases, scanning those paper batches, and securely emailing batched lab results to public health nurses to process.

After all the paper records were scanned, batched and securely routed to the public health nurses, the paper records were then routed to a data entry team to enter into the county’s case management system.

The initial retrieval of the lab results required one or two public health nurses, who along with administrative support printed, collated, scanned and distributed the lab results. This typically took about four hours. A team of data entry staff, ranging from six to 12 employees, were tasked with manually entering the positive test cases, from the paper lab results, into the case management system.

[ Don't miss: Mike Elgan every week on Insider Pro ]

This could take as much as 12 hours to complete, and involved employees working multiple shifts.

When the number of positive cases received on a daily basis from New York State exceeded 1,000, “the process became unmanageable and we were typically scrambling to find employees who could be added to the data entry team for that particular day,” Mastellon said.

In addition, the number of hours worked had an impact on the quality of the results, Mastellon said. On March 20, the first day the IT department engaged UIPath, the county had a total of 459 positive cases. Within a week, it had 3,385 cases and was averaging close to 400 positive cases per day.

The county implemented an RPA robot to automatically retrieve test results from the state and load them into a secure database within the county. All attribute information available from the state heath system were pushed to the secure database.

The automation continued as the robot then selected positive cases, logged into the county’s case management system and added them automatically. Now, test results are retrieved every 30 minutes from the state system, loaded in the secure database, and loaded into the county’s case management software.

Suffolk implemented the other two robots to perform the work as the total number of tests increased. “We implemented the RPA robots on April 1, and after one week of quality review were able to reassign our data entry team to other value-added activities in our COVID-19 response efforts,” Mastellon said.

Once all positive cases were being automatically entered into the case management system, the county was able to leverage the system to electronically assign cases to nurses without having to print, batch, scan and email paper records to public health nurse staff.

The county averaged about 1,025 case per day in the 22 days following the RPA implementation. Mastellon estimates the manual data entry work needed would have taken about 1,500 hours. This equates to about $33,000 in pay for just three weeks, he said.

Over a three-week period during the height of the pandemic in the area, the county loaded 75,973 lab results into its secure database, and entered 24,799 positive cases into its case management system.

“Since having the positive cases automatically loaded in our system, we have had the ability to assign cases for patient interviews on the same day, and reduced our time to contact to approximately one day,” Mastellon said. “Previously, since everything was in paper, we were unable to determine how long it was taking to initially reach out to the patients.”

By reaching out to patients as soon as possible, the county was able to educate them on what do to and how to avoid the spread of the disease. “We feel that this has had a significant impact on reducing the overall spread within the county,” Mastellon said. 

Given the success achieved with the RPA project and the ease by which the county was able to implement the system, “we will absolutely look to expand our use of RPA within the health department and many other departments within the county,” Mastellon said.

While the tragedy of COVID-19 will go down as one of the worst in the county’s history, “if I was to take away one positive from this tragedy, it would be that it has challenged us to develop innovative solutions using technologies such as RPA,” Mastellon said. “That will only benefit the county and our government in the long-term.”

Related: Connected RPA Delivers A Digital Workforce

Best practices for effectively deploying RPA

Organizations planning RPA implementations would be wise to create a center of excellence (CoE) and a “clear litmus test” for what to automate, said Frances Karamouzis, research vice president and distinguished analyst at Gartner.

The CoE would cover areas such as understanding existing business processes and tasks, identifying where RPA could help, testing automation tools, and redesigning the structure of a task to be automated.

It’s a good practice to use RPA in conjunction with other technologies, Karamouzis said. Many of Gartner’s clients are deploying RPA and other automation tools in order to make the shift to what the firm is calling “default digital.”

Related: 4 steps to RPA success

“The important point is that very few clients are currently using RPA alone; it’s often used in what we call ‘complemented RPA,’” Karamouzis said. “It’s basically complementary technologies such as machine learning, analytics, process mining, UX/CX [user experience/customer experience], ingestion engines.”

Many enterprises have needed to move their employees to be fully digital in a work-from-home environment, Karamouzis said. “They needed to automate as much as possible as fast as possible,” she said. “RPA is so ubiquitous that it can be used in tons of tasks and processes.”

If the idea is to get automated processes up and running quickly to address an immediate need, organizations don’t need to be overly focused on long-term goals.

“Quickly putting up a software robot is not necessarily focused on long-term resilience and manageability,” Fleming said. “The job is primarily focused on just getting it done. Once in production, the software robot may be improved to increase throughput. In this case, building a viable but less efficient robot makes the most sense.”

Performance improvements, a focus on resilience and ongoing change management can be added incrementally, Fleming said.

“This is easier when the enterprise and team have a program and skills in place with existing conventions, governance, and pre-automated elements that can be readily leveraged [in] an emergency,” Fleming said. “Front-end automation that controls interactions with a library of application UIs [user interfaces] and pre-built capabilities that access back-end applications can be helpful, and it is far more likely that the robots will be built well to begin with.”

To effectively deploy and maintain RPA platforms on an ongoing basis, companies should make efforts to simplify and standardize processes before expecting an automation payoff, Le Claire said.

“Humans just find too many ways to get the job done,” Le Claire said. “Different short cuts, tools, or task order are adopted, even when the desired outcomes are the same. So before RPA automation, understand where variation in a task can be reduced through training and elimination of steps.”

Keep in mind that people are critical to the success of automation, Le Claire said. “Human insight, labor, and support is often required to plan for, scope, deploy, and stabilize automations,” he said. “Straight-through processing is just not viable in all use cases. Scenarios where bots and humans interact need to be carefully designed.”

In addition, companies need to align RPA automation with the right use cases.

“Technology teams are eager to experiment with RPA and often forget the adage, ‘to a hammer, everything looks like a nail,’” Le Claire said. “Other tools within the intelligent automation toolbox may achieve a more durable result, or with a path to greater intelligence.”