HSBC plans self-learning RPA expansion

HSBC has been deploying robotic process automation (RPA) since 2016 to eliminate manual processing. In 2019 it brought its RPA capabilities in-house, rather than rely on external vendors.

Now it is planning to deploy RPA-plus, adding elements of self-learning to the software so that it can tackle even bigger problems, says a bank executive involved with HSBC’s global transformation team.

The job of the operations center is to save the bank money by using technology to eliminate headcount. It uses the term FTE, or “full-time employee”, as its unit of measurement. In the past year the center has notched close to 200 FTE “saves” in China, and far more globally.

A bank can reduce headcount by training operations people how to automate the tasks they would normally require I.T. people to resolve. In HSBC’s case more than 170 operations people in China have been trained to use RPA tools so far, and learned they can digitize their workflows by themselves.

“We have confirmation that the FTE is demised,” said the source, jargon for full-time employees. In other words the bank is able to cut, not reshuffle, costs.

I.T. and ops

Banks have been automating processes for years, especially at the infrastructure level: the plumbing, connecting systems and databases. But there have remained “exceptions processing”, often the smaller tasks that fall between the cracks, and which have been too minor for I.T. teams to bother addressing. But added up, exceptions processing represents a towering amount of dull but necessary work.

RPA automates repetitive human activity at the front end, that is, where a human interacts with the automated plumbing. At its most basic level, RPA is a rules-bound algorithm designed to handle specific tasks. It is designed to be relatively simple, like an app on your mobile phone – simple enough that a non-tech person can learn how to use it.

Although HSBC is deploying RPA to shed excess I.T. headcount, the use of RPA also frees its core I.T. teams to focus on big, transformative projects, instead of having to help operations teams sort out pesky workflows.

RPA plus

Lee says the bank now rolled out RPA solutions across all its China operations teams supporting commercial bank products, retail bank products, KYC, onboarding, and other functions. The digital operations center in China also supports internal businesses in Japan, Malaysia, and other markets in the region.

Now the bank is adding analytics to its RPA suite. “This isn’t just filling in the blanks of a form, but adding another layer of intelligence,” said the bank executive. This includes work such as optical character recognition (OCR, how machines read text) and indexing or categorizing data.

In practical terms this will allow functions in artificial intelligence such as voice-to-text conversions or analyzing text to understand the nuance of a conversation. The business goal is to apply it to call centers, starting in 2021.

Blue Prism’s software

Blue Prism, a U.K.-based RPA software and integration vendor, has been working with HSBC since 2016. The bank also relied on consultants such as EY to run its first RPA solutions, before bringing that in-house last year. But it relies on Blue Prism’s software support, upgrades, and analytics.

“We have a strong partnership with Blue Prism,” the banker said. “They’re bringing in an intelligence aspect.”

Dan Ternes, Blue Prism’s Hong Kong-based CTO for Asia Pacific, says HSBC has been adept at taking full advantage of what RPA can offer. “They’re not treating it as just a shadow I.T. department,” he said.

Although some enterprises dip a toe into the RPA waters by applying it to simple problems that may be isolated from the broader business operation, Ternes says such tactical uses can be self-defeating. “It’s easier to implement the first time in operations without too many moving parts, but then you can’t repeat it at scale,” Ternes said.

HSBC’s senior operations and I.T. execs, on the other hand, have made RPA a strategic business tool.