- Key insights: Bank of America is connecting higher revenue and lower expenses to the use of AI.
- What’s going wrong: The success of AI implementation will determine operating costs and competitive positioning.
- Support data: Erica has managed 3 billion customer interactions, effectively replacing around 11,000 employees.
Bank of America’s Erica, a virtual assistant, is currently handling tasks equivalent to those done by 11,000 call center agents. Chief Technology and Information Officer Hari Gopalkrishnan highlighted this at a recent event as an example of how banks are benefiting financially from artificial intelligence.
The bank has a technology budget of $13 billion, with $4 billion allocated specifically for new initiatives. During an interview at the Reuters AI Momentum conference in New York, Gopalkrishnan told finance editor Ranan Nguyen that the bank’s technological focus is on enhancing customer experience, boosting revenue, and cutting costs.
His comments come amid a broader landscape where many companies are yet to see returns on their AI investments. A study from MIT found that, as of August, 95% of firms have not gained any benefit from their AI spending.
However, Gopalkrishnan expressed optimism, stating that “we cover eight distinct lines of business, and there’s potential for revenue everywhere.”
For instance, AI is increasing the productivity of relationship bankers. “Consider how much time goes into getting ready for customer meetings,” Gopalkrishnan noted. “If they could automate the prep work with AI and analytics, one banker could serve 50 clients instead of just 15. That’s what we’re witnessing in practice.”
These bankers utilize AI tools for meeting preparation, which allows them to request insights about specific customers. “I’m meeting with customer X tomorrow. Can you prepare everything they might need?” Gopalkrishnan explained. “We gather both structured and unstructured data to deliver real-time insights.”
When it comes to cutting costs, he emphasized that AI can provide clarity on expenditure.
“Every second spent in call center processing translates to significant costs,” he remarked. “By automating predictions for customer inquiries, we can analyze voice-to-text data to better understand needs, enabling agents to respond more efficiently, saving around 30 seconds per call.”
Over the past eight years, Erica has been involved in 3 billion customer interactions.
“We witness the equivalent of 11,000 jobs being done daily through Erica,” Gopalkrishnan noted.
In the past, customers used to call the center for issues like mistaken charges or check orders. “With Erica managing these calls, the need for human agents has decreased dramatically,” he stated.
He also mentioned that AI has been instrumental in analyzing 7 million cases of fraud related to person-to-person payments.
“Every Zelle payment undergoes evaluation through two AI models to verify identity, amount, and the recipient,” Gopalkrishnan explained. “This AI has automated processes that significantly lower fraud-related losses.”
To mitigate potential errors and risks, Bank of America evaluates these models based on sixteen different risk factors, including ethics and privacy concerns.
In the wealth management sector, the bank employs an AI-driven market analytics platform that provides financial advisors with daily updates, which they can then align with their clients’ portfolios.
“This allows advisors to call clients with relevant market insights without needing to collect information from multiple sources,” Gopalkrishnan said, reflecting how much simpler the process has become.
In summary, he suggested that AI is “streamlining efforts throughout the system.” “In three years, we envision a much more effective system, opening doors to even more innovative possibilities ahead.”
