As banks have had to adapt to the majority of their staff working remotely, they are increasingly embracing automation, and accelerating their digitization journeys. One technology that has enabled this is robotic process automation (RPA)—although many skeptics remain. 

The issue with RPA, especially if interwoven or layered with legacy technology, is that it can be challenging to maintain.  

Vivek Veeraraghavan, vice president for process optimization for Asia-Pacific at Northern Trust, said the industry looks for technologies that will guarantee “quick wins”.

You do not want to automate a bad process. You want to ensure that your automation is worthwhile, because the ongoing maintenance is quite high

Vivek Veeraraghavan

“Especially for us in the banking industry, it’s quite important to know what we are going to automate. Yes, all of us want to leverage the technologies; we want to really get quick wins. And, under the current circumstances with Covid, every organization is under cost pressure,” he said, speaking at the WatersTechnology Innovation Exchange. 

That said, firms should not “jump into” automation, just because a process is rule-based and can be automated. Vivek said. Instead, firms should carefully consider the appropriate solution to improve the process.

“A simple process change could lead to [better] efficiency. So there is a compelling need for us to take a step back and assess whether the automation really needs to be done. You do not want to automate a bad process. You want to ensure that your automation is worthwhile, because the ongoing maintenance is quite high, and the fact of the matter is you will have to have maintenance in terms of production support type of work,” he said. 

Firms should be especially aware of this when looking to automate processes that have specific “cut-off” timeframes, such as in settlement. 

Vivek added that in the current situation, there is an accelerated need to look at RPA solutions. “Quick wins are important. But, ‘at what cost,’ is the question. We need to really take a step back, assess and then go about implementing, [and] maintaining some of the bots,” Vivek said. 

David Sharatt, China COO at Commerzbank, who spoke on the same panel, said it’s important to question if RPA—or any other technology, for that matter—is the right solution to improve any particular process. Rather, he believes that business will derive the most value for investment from sophisticated applications like machine learning or intelligent solutions, as the tech continues to evolve.

“I think ever more [apparent] in the current environment—commercially and operationally—is [this idea of doing] RPA for RPA’s sake, [and it] is really wasting your resources, because the who, what, why, and when are really the valid questions that you have to ask. Is there quite simply an alternative way to do what I’m currently doing?”

Commerzbank is currently looking at wholesale re-evaluation of its process chain, including the way it executes, documents, and settles transactions. Post-execution, in particular, is one area ripe for automation, Sharatt said, and as regulations continue to change, the bank will also look at other third-party offerings to apply more sophisticated AI. Sharatt said he only sees value in using RPA on more smaller-scale use-cases.

“There is the ability to put RPA into the execution space as well, but I do think the lens is now going to shift very much to [asking]: ‘Am I doing the right thing, in the right place, with the right technology, and should I even be doing it in the first place?’”

RPA can be useful as a quick fix, and to solve immediate requirements, added Venkatesh Subbaraman, regional chief operating officer at ANZ. RPA can also be used as a springboard to other technologies, such as machine learning and distributed ledger technologies, he said, but controls need to be in place in place for the bots, just as they are for human employees.

“We need to put controls around the bot,” he said. “We have had situations of bots misfiring and not doing their job. Once you do that, then you can look at the next [technologies] you’re planning [to use], so on and so forth. It’s a really promising toolkit that will actually help us solve real-life problems.”