Banks Are Just Beginning to Understand the Power (and Limitations) of Artificial Intelligence

Citi Ventures‘s recent investment in Anaconda, an Austin, Texas-based AI and machine learning (ML) software company, shows the industry’s commitment to staying at the forefront of this growing and developing field.

AI has certainly generated a lot of chatter and buzz in the financial world, but banks are still challenged to implement it properly.

Citi Ventures’s co-head of venture investment, Ramneek Gupta, told Bank Innovation he thinks these problems mainly stem from finding the right talent to manage machine learning operations, building necessary computing power, getting products into production, and the way that banks are siloed.

“A place a lot of financial services institutions have challenges is creating access around the various silos that exist within the bank,” said Gupta.

But he is confident that banks can get more out of their propriety data to use in AI applications.

“Where banks once used statistical modeling capabilities, banks are now levering ML capabilities,” he said.  “AI is an evolutional trend, in the number of applications, which are already very data forward and data focused.”

It’s this evolution of AI that Citi has invested in, Gupta says the investment in Anaconda was years in the making.

“We’ve been tracking Anaconda for the better part of two years,” said Gupta. “In that period, we helped them get access to Citi, and the capabilities of the platform, that culminated in adoption into Citi’s architecture.”

Accessing Expectations 

Anaconda CEO Scott Collison told Bank Innovation that some of the more common use cases among the banks that use their platform, are AML, realtime credit decisions, and modeling for stress testing. Some of the banks currently utilizing the platform are JPMorgan Chase, Goldman Sachs, Bank of America, Merrill Lynch, HSBC, and others.

“It’s very hard for a bank to write a general-purpose platform for enabling AI,” said Collison. “Banks are good at writing single-purpose applications, online banking, trading applications, but if you’re going to run a variety of things NLP, anti-fraud, AML, real time credit decisions, that forces you to write a general-purpose platform.”

The current use cases for AI in banks may have also set the bar of expectations too low.

“There were some misplaced expectations and people thought that as long as we have natural language processing (NLP), for example, then we have solved the problem of chatbots, but it’s only the beginning,” Eran Livneh, vice president of marketing at Personetics, a AI and banking software provider, told Bank Innovation.

“The main advantage AI can bring in terms of customer experience is helping customers manage their day-to-day finances better,” said Livneh.

4 - Readers Like This Post