Finance is the ideal place to integrate robotic process automation (RPA) into your company's processes and free up staff for higher-value work. But IT should be part of the integration from the start, RPA experts said. 

The lion's share of RPA integrations today — about 85% — are led by the finance team — the appropriate place to start because discrete parts of a company's accounting and finance functions are ideal for automation, said Justin Gillespie, principal of digital technology services and solutions at the Hackett Group.

Gillespie discussed RPA adoption in a CFO webinar released last week. About 10% of RPA integrations are led by other financial areas in a company — HR or supply chain, typically — and about 5% are led by IT. 

It's best to have finance lead the integration because no other function area has more opportunity to test automation on repetitive, manual tasks, said Gillespie and Hector Barletta, senior director of strategy and business transformation at Hackett. Finance is also best positioned to understand how to integrate automation in a way that works across divisions, they said.

"You need someone watching the house," said Barletta, referring to the need to create a finance-led team that can ensure automated processes don't cause problems in other areas of the company.

Although finance should lead the effort, it should only do so with other teams as part of a Center of Excellence (COE) or other type of integration group, and IT must be part of the team to ensure it works with security protocols and other technology the company uses. "It often starts with finance, but it might not end up in finance [after it scales]," Barletta said.

Integration starts with scans

The integration experts suggest finance chiefs scan their operations and identify distinct areas that can be automated as pilot programs and then expand robotic processing over time to more areas, eventually crossing over into HR and other areas.

Gillespie recommends CFOs look at five dimensions when deciding where to start adding automation:

• Process frequency. How often is a task undertaken? Hourly? Daily? Weekly?
• Type of input. Is the information digital? Can it be made digital?
• Rules-based. Can the operation be reduced to a set of rules?
• Process stability. How long has the process been in place? Has it proved workable, or is it subject to change?
• Volume. How much data is being processed?

The perfect test case for adding RPA is manual journal entries, Barletta said. If you have staff typing data into a template and loading it into an enterprise resource planning system — what Barletta calls a "swivel chair process" because employees swivel back and forth between material on their desk and templates on their computer — you have a process that can serve as an early test for automation. "You can point the software to read the data where it needs to read it from, even if it's in an email, and enters it into as template, then emails it to a person for review, and upon approval, it can load that journal entry," Barletta said.

Early testing can be done in days and weeks rather than months and years, Gillespie said. Once you prove the concept and start seeing the time savings, you can add RPA to more function areas. Then you start to identify more areas where automation can work, and it balloons from there, he said.

As more processes become automated and finance personnel are freed up from repetitive tasks, they can spend more time on higher-level analytics. "Finance professionals didn't get into that function because they want to cut and paste data," Barletta said.

Once employees see how much time RPA saves them, they'll see its adoption isn't about replacing them but allowing them to make better use of their time, Gillespie said.

"There's a fear around automation that it will take someone's job, but it only takes pieces of a person's time, and by freeing those people to do other things, they can focus more on analytics," he said.

Finance staff are well-positioned to look for more ways to apply automation from there. "Automation begets automation," Barletta said. "It can really snowball within the organization."

What's more, widespread RPA adoption works as a gateway to adopting other value-adding automated processes, including artificial intelligence and machine learning, to extract more value out of your data and processes, Gillespie said.

The additions are relatively seamless because RPA is structured to work with "cognitive" AI and machine learning software, Barletta said.