Milos Maricic: listen for the AI number on next week’s bank calls

On September 14, Bank of America CEO Brian Moynihan announced that the bank had 130 to 140 AI projects in place, costing $400 million but generating $800 million in benefits. The bank had never disclosed such a figure on an earnings call before. When Bank of America released its third-quarter results on October 14, it was unclear whether to expect a mention of the $800 million figure.

A review of 919 earnings calls of 60 major U.S.-listed financial firms since January 2023 found only one firm, S&P Global, providing a specific dollar figure related to AI savings - around $9 million a year from its Kensho Scribe tool. Analysts asked for such information 29 times without receiving a response. Among 100 major U.S. companies outside tech, communications and finance, only FedEx has disclosed AI cost savings in a call, crediting a new AI image capture process for an annualized benefit of over $180 million.

This represents just two companies in 160. Executives have shown reluctance to disclose these figures publicly, with Goldman Sachs CEO David Solomon stating, "We don't have the full confidence to put information out publicly." In June 2025 at a Morgan Stanley conference, JPMorgan Chase CEO Jamie Dimon claimed the company benefits to the tune of $2 billion from AI investment, yet the figure has not been mentioned on any of JPMorgan's 13 earnings calls since mid-2023.

In October 2025, CFO Jeremy Barnum admitted proving AI savings is a "very hard thing to do, hard to prove." The article highlights three reasons for skepticism regarding AI's impact on financial institutions: the calendar, the fading effectiveness of AI, and the high costs. AI investments are expected to yield benefits on the "productivity J-curve," where firms are currently located.

Historical evidence shows a lag between AI adoption and measurable productivity gains. The benefits of AI can also fade quickly, as seen in the decline of China's long-only quant funds that initially outperformed human stock pickers. Moreover, AI costs are becoming a significant concern, with AI cost mentioned on 53% of financial calls in the past season, up from just 3% in late 2022.

Optimizing AI spend can remove about 40% of annual AI expenditures. The article concludes by suggesting investors should pay attention to whether executives provide specific dollar figures for AI savings during upcoming earnings calls. If Bank of America repeats Moynihan's $800 million figure, it would be the first of the 21 banks tracked to disclose such information on a call.