Banks are expected to expand the tracking and disclosure of AI-related financial gains amid growing market interest in the technology's economic value.
AI adoption across the sector accelerated in 2026, with many banks broadening use cases and rolling out generative and agentic tools to staff and clients. As the sector's push to scale AI beyond pilots has gained traction, so have discussions about investments in the technology and the benefits it promises to bring to the bottom line.
More investment demands proof of returns, and banks are likely to provide more details about that in the future, experts told S&P Global Market Intelligence.
"The pressure is coming, and honestly, it's about time," said Nicolas Pinto, vice president of marketing and partnerships at French fintech RollingFunds. "The last couple of years were a bit of a free pass; everyone was piloting things, and boards were patient because the story was exciting enough on its own. That patience is running out."
More banks quantify gains
While measuring the benefits directly linked to AI remains challenging, more banks are speaking about return on investment and some of them are providing concrete figures.
As many as 12 of 50 of the world's largest banks reported realized and projected ROI as of October 2026, up from eight in 2025 and four in 2024, Evident data shows. This kind of reporting is difficult, but important given the heightened scrutiny banks face from clients and shareholders, Alexandra Mousavizadeh, co-founder and co-CEO of Evident, said in an interview with Market Intelligence.
JPMorgan Chase & Co. was among the first banks to speak about investment versus benefits, with CEO Jamie Dimon saying in June 2025 that the US group's roughly $2 billion of investment is matched by about $2 billion in realized benefits.
Bank of America Corp. has implemented about 140 use cases at a cost of $400 million, so far generating $800 million in benefits, CEO Brian Moynihan told a business conference Sept. 14. The US group plans to double its investments in AI in 2027, Moynihan said.
State Street Corp. has projected annual run-rate benefits of roughly $1 billion by 2029 from its digital transformation program, with enterprise-wide AI adoption a key component. It expects about $750 million of the annual benefits to come from productivity improvements and the rest from revenue uplift.
Rise in tracking, disclosures expected
Major European banks have also revealed more details about how they measure cost and revenue gains linked to their AI deployment, with some also disclosing specific figures earlier in 2026.
Lloyds Banking Group PLC expects over £100 million of cost and revenue benefit from deployed generative AI (GenAI) use cases in 2026, compared to the £50 million estimated benefit from about 50 GenAI use cases it deployed in 2025, the UK bank said in late January. BNP Paribas SA estimated in February that its AI-related cost and revenue benefit will increase to €750 million in 2026 from €635 million in 2025.
While many banks report productivity improvements, cost savings, automation rates, and faster delivery times, AI benefits are often embedded within broader transformation programs, making attribution challenging, Morningstar DBRS credit analysts Maria Rivas and Nicola de Caro told Market Intelligence. But as AI deployment becomes more industrialized and investment levels increase, disclosures are likely to become more granular, with banks increasingly linking AI initiatives to specific efficiency and revenue outcomes, they said.
Banks have been expanding their tracking in 2026, increasingly measuring the potential revenue benefits that automation could bring beyond improved client services and experience, said Will Lawrence, CEO of Bretton AI Inc., a San Francisco-based company specializing in automating compliance and risk workflows in the financial sector.
"The first wave of AI has always been measured on cost savings and efficiency," Lawrence said. "But that's almost like measuring the value and impact of a computer on how much faster it can help you type. It's a very low-level basic input. There are so many more revenue-generating opportunities that we're now seeing people calculate in."
Financial institutions are now interested in gauging how much revenue uplift can be derived from faster onboarding of their most valuable clients, for example, Lawrence said. And they are seeking to generate competitive advantages using AI, not just to cut costs, he said. For example, a bank's loan underwriting model would keep improving and would lead to more high-quality lending decisions, Lawrence said.
Payoff forthcoming
Increasing investments in AI are bound to improve cost-to-income ratios, share prices and profits eventually, Evident's Mousavizadeh said during a webinar discussion Oct. 8. It may not show up this year or next, but it surely will in the years to come, she said.
Furthermore, banks must continue investing or risk losing out to the competition, she said.
- + Explore Bank of America's latest calls and presentations on S&P CIQ Capital Pro
- + Read about how AI adoption is transforming bank jobs
|
Banks that focus on the right areas and right workflows are going to see "massive" returns, Aser Blanco, head of banking at NVIDIA Corp., said during the webinar. They could get returns of up to 10x their investment in areas like fraud, for example, where a $20 million spend on building the necessary intelligence could lead to savings of $150 million every year, Blanco said.
As banks move beyond single use cases to enterprise-wide deployment, AI capabilities could be reused across processes, lowering marginal costs and supporting longer-term business model improvements, Bank of New York Mellon Corp. executive Isabel Schmidt said at a recent conference.
Transitioning to a superior technology carries a cost but eventually leads to a "substantially better business model" with increased returns, said Melissa Tuozzolo, global head of client services at HSBC Holdings PLC, during the same panel discussion.
"When you're thinking about the ROI and the investment, it's really important to think about the macro five years from now," she said.
Historically, rising transaction volumes at HSBC meant higher service costs because the group needed more capacity to respond to queries.
"This is the first year … that we're actually seeing it go the other way," Tuozzolo said. While not entirely due to AI, handling queries much faster has helped free up capacity, she said.
