A Bot In Need Is A Friend Indeed

After Hurricane Irma carved a path of destruction along Florida’s Gulf Coast in the fall of 2017,Suncoast Credit Union ($9.2B, Tampa, FL) was flooded with emergency member loan requests. Within hours, the credit union deployed robotic process automation (RPA) to resolve a backlog of nearly 5,000 applications.

Suncoast is the largest credit union in the state of Florida, with nearly $9.2 billion in assets and more than 792,000 members. As a state community-chartered cooperative, Suncoast serves 21 counties along the Gulf Coast through its network of 65 branches.

Category 4 Hurricane Irma made landfall in the Florida Keys on Sept. 10, 2017, and rapidly moved northward up the coast, almost perfectly in line with Suncoast’s branch footprint. Recognizing the potentially devastating impact on the credit union’s membership, Suncoast’s senior management activated its Lifeline loan. Lifeline is an emergency personal loan with special terms, a low monthly payment, and a low interest rate that Suncoast deploys in the case of natural disasters, government shutdowns, and other disruptions that prevent members from accessing cash for short-term needs.

Suncoast accepts consumer loan applications through its mobile banking app, among other channels. In the days following Hurricane Irma, the credit union posted a message on the front page of the mobile app announcing the availability of the loan. Immediately, loan applications jumped from a manageable daily volume of three to four hundred to a high of 4,000 requests in a single day.

Suncoast’s consumer loan origination system is set up to automatically pull credit at the time of application. However, the software has controls in place to limit the total number of credit pulls in a single day, helping to mitigate the risks of cybercrime, fraud, or other unusual and suspicious activity.

Unfortunately, the 4,000 requests Suncoast received were far above its system’s daily credit bureau limit, which created a backlog of unprocessed application requests. Suncoast’s leadership and IT teams sprang into action to implement an efficient, automated workaround for this critical bottleneck.

Our technology team had been doing some work with robotic process automation. Within maybe three hours, they had the bot up and running to pull the backlogged credit scores for us.

“Our technology team had been doing some work with robotic process automation,” says Vicki Lovett, chief lending officer at Suncoast. “We thought this would be a perfect use case for it. We have really talented programmers, and our leadership team spent some time with them to explain all the loan processing steps. Within maybe three hours, they had the bot up and running to pull the backlogged credit scores for us.”

Since the bot was only set up to perform the menial task of pulling credit scores, it did not impact credit quality or increase the risk in the loan portfolio. Once the bot pulled credit, each application went through the normal underwriting process. Suncoast normally uses automated decisioning models for certain loan types; however, because Lifeline is designed for short-term, emergency purposes, lenders at Suncoast had to put in a bit more work to help those members with credit challenges and the most urgent needs.

By the time Suncoast turned off Lifeline for Irma relief, it had processed more than 9,700 loans for a total of $18.8 million. Roughly 5,000 of those were processed with the help of the credit-requesting bot.

“Our loan operating system normally pulls the credit,” Lovett says. “When we got that huge influx of emergency requests, it quit working. Once we returned to a normal volume of applications, the system began pulling credit again. The robot helped us clear the backlog very quickly.”

Luckily for Suncoast members, the credit union had deployed the first proof of concept for RPA just prior to Hurricane Irma. Suncoast credit and debit cardholders can set travel notifications within the mobile app or online banking to avoid temporary card blockages. Suncoast’s technology team implemented a bot that transmits the information members enter online to the credit union’s card processor, without any human intervention. As Irma stormed through the Atlantic and approached Florida, Suncoast experienced a massive increase in members leaving the state. The bot successfully handled the increase in volume from a few hundred notifications to several thousand over the course of a few days.

Although based in Florida, Suncoast did have members who were impacted by 2018’s Hurricane Florence, which made landfall in eastern North Carolina. There, too, the credit union offered the Lifeline loan to members in need.

“We helped some members in the Carolinas because we put Lifeline in place for people who live in hurricane-impacted areas,” Lovett says.

Lovett views the test of RPA technology as a success story, and Suncoast has since deployed the technology in other operational areas.

“We are continuing to build out RPA at Suncoast in a variety of our business units”, Lovett says. “It works particularly well for repetitive processes that require you to perform the same tasks over and over again.”

That might sound like dull work, but what’s not dull is the opportunity Suncoast sees in bots.

“We had set a goal of replacing 8,000 human hours with bot hours by the end of this year,” says Ted Hassenfelt, chief information officer at Suncoast. “We are focused on simple, menial processes like loan documentation and archiving. The idea was not to eliminate positions, but to allow team members to focus on other tasks. Already, we have surpassed our goal and have replaced more than 10,000 human hours with bots, and by the end of the year we anticipate hitting 16 or 17,000 hours.”

Overall, the RPA experiment has worked out well for Suncoast.

“I talk about RPA with five to six credit union CIOs every month, and I encourage them to employ bots in their organizations,” Hassenfelt says. “RPA is quick to implement, and it delivers an unbelievable return on investment. Not everything is appropriate for a bot, but there are many processes with gaps that a bot can fill.”

Indeed, these bots can be a friend in need.

“Our IT team was already investigating the RPA capability and thought we could do this,” Lovett says. “Sure enough, we did. It helped us serve our members in their time of need, and I think it worked out really well.”