Bank of America logs 400,000 AI prompts daily, 114 live use cases

Bank of America disclosed substantial scale in its AI deployment during its second quarter earnings call, with CEO Brian Moynihan and CFO Alastair Borthwick providing updates on the breadth and impact of the program. Across the bank's 200,000-person workforce, employees are generating more than 400,000 AI prompts daily. The approval pipeline includes over 300 AI use cases, of which 114 are live generative AI deployments. Of those live cases, 34 are fully implemented, and leadership stated that new capabilities are rolling out on a weekly basis.

The approved use cases span multiple business functions. They include generating background information and research materials to help customer relationship managers prepare for client meetings, automating the creation of presentations, and enhancing developer productivity through AI-assisted code generation. These applications reflect a mix of internal efficiency plays (faster analysis, better-prepared staff) and customer-facing value creation (improved service delivery and decision quality).

The financial results show the downstream impact. Bank of America reported second quarter net income of $9.1 billion（約1.5兆円）, a 27% increase from the same quarter a year prior. Revenue climbed 15% year-over-year to $31.6 billion（約5.1兆円）. Most tellingly, the efficiency ratio—a key measure of operational cost relative to revenue—improved by 359 basis points to 59%, indicating that the bank is generating revenue with proportionally lower non-interest expenses. CFO Borthwick attributed gains in service, efficiency, and employee focus on high-value work to AI capabilities, and highlighted the scale of digital engagement: roughly 50 million active digital users, more than 24 million active users of Erica (the bank's digital assistant), and digital sales representing 70% of total sales.

When pressed on return on investment and future spending, leadership offered cautious optimism without detailed financial models. Borthwick noted "growth and margin improvement" but acknowledged the bank is still "working our way through" the investment landscape. Moynihan elaborated that the bank is spending "at a good clip" on technology and dedicating internal effort to "careful examination, implementation, catalyst, and people working to understand the projects in AI." He pointed to productivity increases but stopped short of quantifying them. He did note that "the coding process has become more and more efficient using these tools," implying that the same technology budget in 2027 will generate more code output in 2028. Overall, he said, the bank is focused on operating leverage and has managed to maintain efficiency gains despite ongoing AI investment.

Bank of America's disclosures on its second quarter earnings call underscore how large financial institutions are moving AI from pilot to production at scale. With 200,000 employees generating over 400,000 prompts daily and more than 300 approved use cases in the approval pipeline, the bank is treating AI as both a cost-reduction and a revenue-enablement tool. CEO Brian Moynihan's remark that "new capabilities coming on every week" suggests a continuous deployment cycle rather than a one-time rollout, indicating organizational commitment to iterative AI expansion.

The financial results validate the productivity thesis. The 27% year-over-year jump in net income to $9.1 billion（約1.5兆円） and the 359 basis point improvement in the efficiency ratio—which measures non-interest expense relative to revenue—point to measurable operational gains. CFO Alastair Borthwick connected these wins explicitly to AI, noting improved service delivery and the ability for employees to redirect effort toward higher-value client interactions. The bank's digital ecosystem (50 million active digital users, 24 million Erica users, 70% of sales flowing through digital channels) provides the infrastructure through which many of these AI capabilities reach customers and generate revenue.

However, the bank has not quantified specific return on investment for individual projects, and leadership indicated continued uncertainty about long-term technology spending trajectories. Moynihan's observation that "the same amount of money in '27 will get us more code…in '28" reflects confidence in developer productivity tools but stops short of a detailed cost-benefit model. Bank of America is positioning itself on both sides of the AI wave: financing infrastructure buildout while capturing internal efficiency gains—a diversified play that reduces dependency on any single outcome.

No comments yet. Be the first to share your thoughts!