Hauling legacy systems to the new age: strategies that work for big banks

Earlier this month, Tearsheet hosted its much-awaited The Big Bank Theory Conference, where minds from the likes of Bank of America, Chase, and Citi sat down to discuss how they have continued to innovate despite the turbulence of the past few years.

One thing that seems to have worked out in the pandemic years is the engagement of audiences that were previously reluctant to use digital channels to manage their finances. While Gen Z and millennials had been more open to digital banking, older demographics remained loyal to their tried-and-tested local bank branch, until the pandemic made it impossible to access.

“Engaging a demographic that has historically not been as digitally engaged and is now sticking has been transformational for us,” said Hari Gopalkrishnan, Head of Retail, Preferred, Small Business & Wealth Management Technology at Bank of America.

This naturally opened a lot more doors of opportunity, but with the caveat that behemoths like Bank of America, Chase, and Citi now had to compete with the pace of change adopted by new fintechs in the market. This is challenging on multiple fronts, but perhaps the most daunting is the scale of transformation needed to innovate at such a pace.

Technologies that had remained in silos for decades now had to be adapted to interfaces that provide seamless end-to-end consumer experiences. As daunting as the task is, nearly every leading financial institution is putting its digital transformation foot forward. Let’s review some of the strategies that are pushing the needle.

Looking under the hood: tech strategies that work

Simpler is better: One possible route to transformation is to adapt current and legacy infrastructures and processes to new products, demands and regulations. As Citi’s Global Head of Technology, Nick Nadgauda explains, this process can lead to a complex infrastructure that is not responsive to the pace of change of today’s digital finance. 

“Not everything needs to scale. Banks struggle at scale because they try and do everything the way they do it today – whether it’s compliance process or risk, they try and scale that entire ecosystem,” he said. “If you're able to simplify some of that and have a different compliance model, then you don't need to talk to the compliance system, the compliance system doesn't need to scale in the same way that it did in the past.”

The crux of the strategy for Citi is simplification. In Nick’s words, “how easy is it to do business with you?”

Align teams to products: Chase has 100 different product teams. Organizations with this level of sprawl in revenue streams can be particularly hard to digitally “transform”. Yet, as Roman Eisenberg, Head of Technology at Chase explains, the key was to “move teams away from projects” and towards products. 

“Focusing on modularization, separating frameworks to APIs, building a testing model that allows product teams to test a lot and pass their features independently. Four hundred or more engineers work independently, they deliver the journey independently for their customers, which opens up tremendous scale in the system,” Eisenberg said.

Experimentation: When organizations develop with agility, products have to roll out quickly, and so do solutions to any problems that arise. This breakneck speed can test the best of teams. One way to mediate this can be through independent experimentation and rigorous testing at the team level. “What we see is that without experimentation, our best intended experiences may result in flat adoption and flat numbers,” Eisenberg added.

Blending high tech and high touch: Chatbots are commonplace now, and even making headlines as ChatGPT by OpenAI captures the modern imagination about automation and the future of work. But in reality, digitizing an agent’s work has proven to be difficult, as Tearsheet has previously covered.

Bank of America’s answer to this problem is ERICA, a chatbot that has served over a billion responses. Here the system allows the user to switch to an agent when they please, but instead of making them call one on the phone, the chatbot includes the agent into the chat. Once consumers are finished with the human agent, they can continue with the chat bot, without ever leaving the screen.