By December 2025, AI adoption on Wall Street had moved past experiments inside large US banks and into everyday operations. Speaking at a Goldman Sachs financial-services conference in New York on 9 December, bank executives described AI—particularly generative AI—as an operational upgrade already lifting productivity across engineering, operations, and customer service.
The same discussion also surfaced a harder reality. If banks can produce more with the same teams, some roles may no longer be required at current levels once demand stabilises.
How Wall Street banks say AI is delivering results today
JPMorgan: operational gains begin to compound
Marianne Lake, chief executive of consumer and community banking at JPMorgan, said productivity in areas using AI has risen to around 6%, up from roughly 3% before deployment. She added that operations roles could eventually see productivity gains of 40% to 50% as AI becomes part of routine work.
Those gains rest on deliberate choices rather than broad experimentation. JPMorgan has focused on secure internal access to large language models, targeted changes to workflows, and tight controls on how data is used. The bank has described its internal “LLM Suite” as a controlled setting where staff can draft and summarise content using large language models.
Wells Fargo: output rising ahead of staffing changes
Wells Fargo CEO Charlie Scharf said the bank has not reduced headcount because of AI so far, but noted that it is “getting a lot more done.” He said management expects to find areas where fewer people are needed as productivity improves.
In comments reported the same day, Scharf said the bank’s internal budgets already point to a smaller workforce by 2026, even before factoring in AI’s full impact. He also flagged higher severance costs, suggesting preparations for future adjustments are under way.
PNC: AI speeds up a long-running shift
PNC CEO Bill Demchak positioned AI as an accelerator rather than a new direction. He said the bank’s headcount has stayed largely flat for about a decade, even as the business expanded. That stability, he said, came from automation and branch optimisation, with AI likely to push the trend further.
Citigroup: gains in software and customer support
Citi’s incoming CFO Gonzalo Luchetti said the bank has recorded a 9% productivity improvement in software development. That mirrors a broader pattern across large firms adopting AI copilots to support coding work.
He also pointed to two customer service areas where AI is helping: improving self-service so fewer calls reach agents, and supporting agents in real time when customers do need to speak with a person.
Goldman Sachs: workflow changes paired with hiring restraint
According to Reuters, Goldman Sachs’ internal “OneGS 3.0” programme has focused on using AI to improve sales processes and client onboarding. It has also targeted process-heavy functions such as lending workflows, regulatory reporting, and vendor management.
These changes are unfolding alongside job cuts and a slower pace of hiring, linking workflow redesign directly to staffing decisions.