Financial institutions are looking to AI to create tools for coding, marketing and customer experience. However, use of AI by banks is being monitored as risk remains top of mind for bank executives. 

Banks are deploying AI throughout their IT infrastructures to make operations more efficient and cost effective, but they are cautious when deploying it for consumer-facing operations to avoid mistakes created by AI, Moutusi Sau, an analyst at Gartner, told Bank Automation News.

Using AI to streamline operations and increase efficiency can be helpful but consumer engagement can be tricky, Sau said.  

“There could be major issues if they get it wrong, like hallucinations, saying wrong things with confidence, and it is not something that anyone wants to do, especially with their brand name tied to it,” she said. 

Back-end AI 

As generative AI continues to raise questions, banks appear to be more confident in implementing the technology within their back-end operations. 

BNY Mellon’s technology arm, Pershing X, is working on an AI-driven platform for strategic asset allocation management, which can help advisers create financial plans for clients, Bill Pierce, director of the Pershing X product, said during Fintech Connect in New York last week. 

The tool uses AI to build a financial plan for a client in minutes, based on data provided to the model, Pierce said

“From a trading perspective, obviously, we want [clients] calling a human if they have a question on trade, not relying on AI for that,” Pierce said. “But within our apps within our operating system, we are looking at having a conversational AI component where the user can use that to get questions answered, as opposed to having to call or sending emails.” 

Nubank Mexico is using AI within engineering sectors and coding, Angelique Wynants, head of Mexico customer experience, said at Fintech Connect.  

Maintaining the human element 

While AI can help streamline operations, front-end capabilities require more exploration, Citizens Bank Chief Information Officer Michael Ruttledge told BAN.  

The Providence, R.I.-based bank is looking to combine AI with human resources, rather than going all in on the tech, Ruttledge said. AI can provide advice or insights for a customer query, but humans help interpret the response. 

Barclays, too, is using AI as a tool rather than a decision maker, Jennifer Warren, head of digital markets strategy, said at Fintech Connect. 

“I think you can’t really remove human oversight from the process right now,” Warren said. “Human judgment has to be a layer on top to ensure that you’re taking everything into consideration,” when making a decision. 

Keeping AI in check 

Whether being applied in the front end or back end, AI needs to be implemented responsibly. 

“If it gives you the wrong advice, there could be many implications for that, and this is what I think everyone wants to avoid,” Gartner’s Sau said.  

However, when implemented fully, AI could deliver value equal to an additional $200 billion to $340 billion annually for the banking industry, according to a McKinsey & Co. report published last month. 

A cautious implementation of AI can help banks to increase productivity and spend their productivity to either double down on their customer base or enter new businesses, Carlo Giovine, a partner at QuantumBlack, McKinsey & Co.’s artificial intelligence arm, told BAN.