As finance tools continue to evolve, robotics process automation (RPA) is becoming the latest addition to financial technology platforms.
At a meeting of FEI’s Committee on Finance and IT (CFIT), Todd Oulton, VP of financial integrity initiatives for American Express, said RPA software operates, at a basic level, like a high-end macro to automate financial tasks or processes by using tools known as robots or “bots.” The bots perform repetitive functions such as downloading data from a transactional system or website and transferring the same data into an ERP system or other platform.
“With RPA, automation mimics human actions,” Oulton said. “RPA can run a report that clears time for deeper analysis. Nobody likes to plug in data from one screen to another, so that’s better suited for RPA.”
RPA is being tested and implemented by a growing number of companies for routine processes such as travel and expense management, invoice verification and payment, credit applications and other back-office procedures.
“Within enterprise IT, RPA isn’t going to replace most technologies, but instead is being used to supplement and connect technologies,” Oulton said. “RPA sits alongside your existing architecture. If you have two platforms that aren’t easy to connect, a robot can bridge that gap.”
RPA uses logic-based rules to process transactions and reports, and to identify exceptions that can be forwarded to a human for further review. This automated precision can help improve control environments by reducing manual errors, and by creating an audit trail of the work the bots performed.
“If your process for paying invoices, for example, calls for someone to sign into your bank to get information, then re-key that information into a spreadsheet and a journal entry, a bot is good for that,” Oulton said.
Firms are implementing RPA for a number of potential benefits, starting with improvements in costs, controls and operating efficiency. Once a bot is deployed, it can operate around the clock without taking breaks or making mistakes (such as transposing digits) that are possible with human labor.
“You want to stop doing things of lower value, and look for higher-value tasks your team should be performing,” Oulton said. “When we fix processes now, we look for ways to tie in automation to make a process better and easier to carry out. We don’t want to automate a bad process and retain it.”
RPA is best suited today for functions or processes that rely on manipulating or reporting structured data, but the technology is expected to become more sophisticated as machine learning and artificial intelligence are adopted by enterprises over the next few years.
While firms have long tried to capture efficiencies by shifting tasks and processes to lower-cost countries, Oulton said RPA offers advantages over that approach by eliminating the need to replace and re-train workers in nations with high attrition rates.
RPA can also help extend the life of older legacy systems by eliminating the need to upgrade those systems or their interfaces to promote compatibility with newer systems. Inefficiencies or quirks that make older systems frustrating for a human can be handled by bots without complaint.
“Any process that relies on people signing into a screen to get information, that’s your sweet spot [for RPA],” Oulton said.