The University of Texas System continues to grow in many ways. Not only is the system’s student population on the rise, but its medical facilities serve more patients each year, its researchers are making more discoveries, and its staff and facilities are increasing the number of services offered to local communities.
“Our system’s missions are to educate students, treat patients, make new research discoveries, and serve local communities—and we’re growing in all those areas,” says Richard St. Onge, associate vice chancellor of the University of Texas System. “All those goals require administrative support, but that’s the area that won’t grow much, if at all.”
With 14 institutions and an operating budget of $21.1 billion, the UT System is one of the largest public university systems in the nation. As the UT System adds more students and more services, there’s an understanding that it won’t add administrative staff or resources, St. Onge says. “We have to figure out how we can do more with the same. And automation is a big answer to that question.”
For the UT System, robotic process automation (RPA) is a new method of automating business office tasks. It is a low-cost, low-risk approach that accomplishes more in less time and with fewer staffers—and it’s easy to see quick results.
Most colleges and universities are in the same boat. Whether the institution is growing or shrinking, the business office is tasked with doing more with fewer (or the same) resources. RPA holds promise for higher education institutions as they tackle this increasingly common scenario.
What Is RPA?
Organizations have been looking for ways to automate processes since the industrial era of the 1920s. As early as 1956, researchers were using the term “artificial intelligence” to describe the capability of computer systems to execute tasks that traditionally require human intelligence and intervention.
RPA is sometimes described as a combination of AI and automation. “RPA is the application of technology allowing employees in a company to configure computer software or a ‘robot’ to reason, collect and extract knowledge, recognize patterns, learn, and adapt to new situations and environments,” according to Intelligent Automation Entering the Business World, a report from Deloitte.
Traditionally, when business officers wanted to automate workplace tasks, the process required copious, detailed work by software developers. To automate one task, developers needed to create a list of individual actions that had to occur. Then they would program the system to complete each action using a computer scripting language or internal application programming interface.
Today, robotic processing automation simplifies this process. RPA tools can repeat a set of demonstrated actions performed by a user, and RPA can be implemented without requiring extra coding or new IT equipment. The computer system can perform tasks the same way every time, reducing errors and speeding up the process. RPA essentially provides a virtual employee, as the software can run unattended, reliably automating repetitive tasks at scale.
Financial industries such as banking and insurance have been using RPA for several years, and the technology is gradually working its way into other sectors. “RPA has been around a long time; heavily regulated industries have been using it for 10 years,” St. Onge says. “But it’s very new to institutions of higher education. We may be on the bleeding edge, certainly on a systemwide level.”
Because RPA works with existing IT systems, it presents a relatively low barrier to entry. When the UT System began using RPA, it worked with Ernst & Young’s Intelligent Automation division.
Building a Strategy
When St. Onge and his team presented RPA to CBOs across the university system in 2017, they established a basic strategy and the standards for using the technology. They also allowed plenty of room for localized decisions and interpretations.
“Part of our strategy is to use the same RPA across the UT System and store automation information centrally so that everybody has access to it,” St. Onge says. “We promote the rules of engagement that CBOs approve, but we want to allow local autonomy.”
The system also established an opportunity assessment process to determine future pipelines and prioritize opportunities. Through ongoing communication and training, CBOs across the system are able to select their own potential projects for RPA. When one campus or office completes a project, all others have access to the details and programming to replicate it.
One major theme in approaching RPA projects is “go get the data,” St. Onge says. Across the university system, hundreds of people spend time each day going to get information from a variety of systems, manipulating that information, and presenting it. And “all that can be automated,” St. Onge says.
Selecting RPA-Ready Processes
When a strategy is in place, the next step is to locate tasks or processes that can translate well to RPA. St. Onge and his team look across large business areas, such as a supply chain or finance, for tasks with some of these characteristics:
- Mundane, tedious, or repetitive operations.
- High-volume transactions.
- Rules-based answers.
- Manual work that is prone to errors.
- Tasks performed during or outside of office hours.
- Existing work supported by algorithms or macros.
When it has a task in mind, the UT System team completes an analytical evaluation of that process. The evaluation includes answering about 20 questions to determine whether the process should be automated, and if so, how much of it should be automated. This usually amounts to 60 to 70 percent—rarely 100 percent—of the task, St. Onge says. Other queries address the value drivers of automating the process in question, such as revenue enhancement, cost savings, cost avoidance, or risk avoidance. This ultimately results in calculating a return on investment, typically over a three-year time frame.
Traditional wisdom holds that organizations should reengineer clunky processes before automating them rather than wasting automation on unnecessary steps. However, CBOs across the UT System have debated this point and some believe it makes sense to automate first and optimize later. “If it’s a clunky process but it’s accurate and automation can run it 20 to 30 times faster than we can do it manually, who cares that it’s clunky? It’s still 30 times faster,” St. Onge says.
Seeing Value
Two years into its commitment to RPA, the UT System now has some significant successes under its belt. When it launched its RPA pilot program in late 2018, the UT System began considering in earnest which tasks would benefit from using RPA. After completing readiness assessments, workshops on operating models, and opportunity assessments, leaders chose to move forward with four of the seven projects they initially considered.
One of the most exciting ways the UT System has implemented RPA is as part of a cash control project at the University of Texas MD Anderson Cancer Center in Houston. The renowned academic institution, cancer treatment facility, and research center attracts a lot of sponsorship and grant monies to help fund the research of various professionals who work there. “There are literally dozens, if not hundreds, of payments coming in every day,” St. Onge says. “Sometimes the information received with the payment is not enough to show where the money belongs—which researcher and which account, for instance. To figure it out, manual processors have to download three or four different files and fill out about 1,000 lines of information.”
Gathering that information and automating the task of directing grant and sponsorship funds to the right research projects and researchers was an ideal project for RPA. Today, commercial software has built the automations and a digital worker runs the automations, documenting them at a keystroke level. This cash control project, which used to take about 100 hours of employee time per month, now takes less than five hours. “Those people are now freed up to do more strategic work,” St. Onge says.
At the University of Texas Health Science Center at Houston (UTHSC-H), another RPA project has relieved department leaders of a time-consuming monthly task. The UT System requires every institution to validate every account on a monthly basis. “It’s probably the most mundane task we have,” St. Onge says. “To get the data, someone from each department must log into systems, download files and queries, find relevant data, and manipulate it” to create a report that’s helpful for end users.
Using RPA, the center has been able to automate that entire process. The digital worker downloads needed files, analyzes the data, and e-mails the contact person in each department, attaching a spreadsheet with the information relevant to that department so they can easily reconcile account transactions. This automated process saves about one workday, or eight hours, per month. “Across a few hundred people with an average salary of $68,750 (including benefits), that generates value (a cost avoidance) of over $250,000 a year,” St. Onge says.
These two successful projects are only part of the story. In the past two years, the UT System has completed 10 RPA readiness assessments and established digital worker connectivity to three campuses. It has also deployed four process automations that will generate $1.1 million in value over the next three years and has completed four opportunity assessments that offer a net value of $4.7 million over the next three years, St. Onge says. Through this process, system leaders have advanced the RPA operating model, and system CBOs are convinced of its value.
More to Come
As the UT System moves forward with RPA, it is with an awareness that it is only the beginning of the UT System’s automation journey. RPA is one of several automation technologies that are expected to transform the workplace, along with machine learning, chatbots, and optical character recognition (OCR). As UT leaders become more familiar with RPA, they’re realizing that it can accomplish even more when combined with some of these other emerging technologies.
“RPA doesn’t learn and doesn’t think, it just performs as programmed,” St. Onge says. “For some tasks, when you need to pull in information from different places, you may need AI to pull together data fields and assemble the information into standard templates and optical character recognition to digitize all of it. Then RPA is gold; it will process the information the same way, every time.”
For instance, each vendor invoice may look a little different, but they all contain the same basic information. Tools like AI and OCR can find the needed information on various types of formats and synthesize it into standard data fields. Once it’s in a standard format, RPA can automatically process invoices, charging each one to the appropriate account with high efficiency and without making errors.
“When we started out, we were only going to focus on RPA,” St. Onge says. “But we’ve now realized we need to adopt some of these other automation technologies sooner than we’d planned, to allow us to maximize RPA.”
As CBOs across the UT System look for more opportunities to implement automation technologies—such as RPA—for repetitious tasks, business officers will be freed up to work strategically, harnessing the full potential of their human ingenuity.
NANCY MANN JACKSON, Huntsville, Ala., covers higher education business issues for Business Officer.