Artificial intelligence is no longer a side project on Wall Street. By late 2025, major U.S. banks have moved beyond pilot programs and are using AI, especially generative AI, as a core part of daily operations. Executives speaking at a Goldman Sachs financial services conference in New York on December 9 described measurable productivity gains across engineering, operations, and customer service.
Those gains come with a more difficult conversation. As banks learn to do more with the same teams, leaders are increasingly open about the possibility that some roles may not be needed at current levels once growth steadies.
How banks say AI is delivering results
JPMorgan: steady gains that build over time
Marianne Lake, chief executive of consumer and community banking at JPMorgan, said AI adoption has already lifted productivity in some areas to about 6%, up from roughly 3% before deployment. Over time, she expects operations roles could see productivity improvements of 40% to 50% as AI becomes part of routine work.
JPMorgan’s approach has been cautious and deliberate. Rather than opening broad access to public tools, the bank built secure internal systems, including its “LLM Suite,” which allows staff to draft and summarize content within tightly controlled environments. The focus has been on targeted workflow changes, clear data controls, and defined use cases.
Wells Fargo: more output before workforce changes
Wells Fargo CEO Charlie Scharf said the bank has not cut headcount due to AI so far. Still, he acknowledged that teams are “getting a lot more done.” As productivity rises, management expects to identify areas where fewer people are needed.
Scharf also noted that internal budget planning already points to a smaller workforce by 2026, even without fully accounting for AI’s impact. Higher severance costs flagged in recent discussions suggest the bank is preparing for adjustments once productivity gains stabilize.
PNC: accelerating a long-term trend
PNC CEO Bill Demchak framed AI as an accelerant rather than a turning point. He said the bank’s headcount has remained largely flat for about a decade, even as the business expanded. That stability was driven by automation and branch optimization, with AI now pushing those efficiencies further.
Citigroup: software and service improvements
Citi’s incoming CFO Gonzalo Luchetti said the bank has recorded a 9% productivity gain in software development, reflecting the growing use of AI copilots to support coding and testing. He also pointed to improvements in customer service, where AI is helping customers resolve issues on their own and supporting agents in real time when human interaction is needed.
Goldman Sachs: workflow redesign and hiring restraint
At Goldman Sachs, AI is being paired directly with changes to how work gets done. Reuters has reported that the firm’s “OneGS 3.0” initiative focuses on improving sales processes, client onboarding, lending workflows, regulatory reporting, and vendor management.
These efforts have coincided with job cuts and a slower pace of hiring, tying AI-driven workflow changes more closely to staffing decisions.
Where AI is having the earliest impact
Across Wall Street, the strongest productivity gains are appearing in work that is document-heavy, repetitive, and governed by clear rules. Generative AI is particularly effective at searching information, summarizing material, drafting content, and moving tasks through approval chains when paired with human oversight.
Early gains are most visible in areas such as operations, software development, customer service, sales support, onboarding, and regulatory reporting. In each case, AI speeds up routine tasks while humans retain responsibility for final decisions.
Governance shapes how fast banks move
For banks, the main constraint on AI adoption is not enthusiasm but control. U.S. regulators already require strict oversight of models, and those standards extend to AI systems. Existing guidance, including SR 11-7 from the Federal Reserve and the Office of the Comptroller of the Currency, sets expectations for model development, validation, and ongoing review.
A 2025 report from the U.S. Government Accountability Office reinforced that traditional model risk management principles apply to AI. In practice, this means banks favor systems that can be audited and traced. Prompts and outputs are logged, performance is monitored, and humans remain accountable for high-impact decisions such as lending and official reporting.
Productivity rises, employment questions follow
The pattern emerging from bank leaders suggests a phased shift. In the first phase, headcount remains stable while output increases as AI tools spread. The second phase begins once those gains are consistent enough to shape staffing plans, through attrition, role redesign, or targeted cuts.
Signals from Wells Fargo’s planning for 2026 suggest some banks are nearing that second stage. Beyond Wall Street, institutions such as the International Monetary Fund and the World Economic Forum have warned that AI could reshape a large share of jobs globally, with outcomes varying by role and region.
What comes next for bank strategy
Banks that benefit most from AI are likely to focus on redesigning workflows rather than simply adding chat tools, strengthening data foundations, and maintaining governance that allows speed without sacrificing trust. The financial incentives are significant. McKinsey estimates generative AI could unlock $200 billion to $340 billion in annual value for the banking sector, largely through productivity gains.
The central question is no longer whether AI works in banking. It is how quickly those gains become standard practice, and how institutions manage the workforce changes that follow.
