Bank of America (NYSE:BAC) and Citigroup (NYSE:C) reported third-quarter 2026 earnings alongside updates on their enterprise AI initiatives.
Citi generated $5.2 billion in net income, up 17% year over year, on revenue of $23.8 billion, an 8% increase.
Bank of America reported approximately $30 billion in revenue and $6.9 billion in net income, including a $1.2 billion regulatory settlement charge. Excluding the charge, net income totaled approximately $8.1 billion, or $1.08 per share.
Technology spending and AI productivity featured prominently in the earnings discussions.
Approximately 14,000 Citi engineers are using an agentic coding tool, helping the bank launch products faster and bring more development work in-house. Technology vendor spending declined 30% year over year, while Bank of America reported a 10% improvement in development efficiency across an operation costing more than $4 billion annually. Its portfolio of approximately 130 implemented AI projects is expected to generate $700 million in annual benefits in 2027.
Why It Matters: AI investments are producing measurable gains in software development and operating efficiency. Citi is tracking faster product delivery and lower vendor costs, while Bank of America is measuring financial returns across its AI projects and improvements in engineering productivity. These results provide benchmarks for evaluating AI investments against technology spending and operational performance.
- Agentic Coding Is Changing How Citi Develops Software: The bank’s use of Devin is allowing engineers to complete more development work internally, reducing dependence on external technology providers. CEO Jane Fraser said AI is helping shorten product development cycles while lowering the cost of bringing new offerings to market. “AI is accelerating our product development life cycle, so we are launching new products at lower cost and at faster speed,” Fraser said. The bank reported a 30%–40% improvement in product launch speed, alongside the reduction in vendor spending discussed earlier.
- Bank of America Is Tying AI Investments to Financial Payback: CEO Brian Moynihan said every implemented AI project has an identified return, with the combined benefits expected to reach two to three times the associated expenses over time. Annual benefits are projected to increase from approximately $250 million-$300 million in 2026 to $700 million in 2027. The bank is also using AI to increase development capacity without a corresponding increase in costs. “We are 10% more efficient, so next year we get 10% more work for the same cost,” Moynihan said.
- Process Automation Is Contributing to Changes in Operating Costs and Staffing: Citi is reengineering approximately 100 manual processes from beginning to end, using AI and automation to reduce expenses and create room for additional business investment. Fraser said the financial benefits generated by these initiatives have already significantly exceeded the bank’s year-to-date spending on AI and automation. Similar efficiency efforts are underway at Bank of America, where process improvements have reduced the need to replace employees in certain functions. The bank reported nearly 3,500 fewer employees year over year while continuing to expand customer-facing teams.
- AI Is Improving Customer Service and Supporting Client Acquisition: Citi reported a 21% improvement in client service response times following the deployment of AI tools for service agents in 82 countries, along with higher customer satisfaction scores. The bank is also piloting Citi Sky in its Wealth business, where AI is being introduced into additional client interactions. At Bank of America, AI-enabled tools such as Erica continue to support customers and employees, while digital channels account for 71% of consumer sales. A generative AI tool launched in September is also intended to make commercial banking prospecting and onboarding more efficient.
- AI Deployment Is Bringing Security and Customer Oversight Into Focus: Technology modernization at Citi has supported wider AI adoption, alongside continued investment in cybersecurity testing and vulnerability detection. Fraser described a security framework for deploying frontier AI models, supported by exercises to improve incident response. At Bank of America, attention has turned to how AI agents interact with financial accounts and the controls governing those transactions. Moynihan emphasized the importance of keeping customers in control of their money, with human authorization remaining essential when AI is involved in financial transactions.
Go Deeper -> Citigroup’s Earnings Report – MarketBeat
Bank of America’s Earnings Report – MarketBeat
Our Latest CIO Field Notes
CarMax Drives Agentic AI Across Its Customer Operations
Stepping on the gas.
