By December 2025, the adoption of AI on Wall Street had moved beyond experimentation within large U.S. banks and permeated daily operations. Speaking at the Goldman Sachs Financial Services Conference in New York on December 9, bank executives described AI, especially generative AI, as an operational upgrade that is already improving productivity across engineering, operations, and customer service.
The same discussion also revealed an even harsher reality. If banks can produce more with the same team, some roles may no longer be needed at current levels once demand stabilizes.
How Wall Street banks say AI is delivering results today
JP Morgan: Operating profits begin to increase further
Marian Lake, chief executive officer of JPMorgan’s consumer and community banking division, said productivity in the AI-enabled area has increased from about 3% before implementation to about 6%. He added that as AI becomes part of daily operations, productivity in operational roles could ultimately increase by 40% to 50%.
These benefits are based on deliberate choices rather than extensive experimentation. JPMorgan focuses on secure internal access to large language models, targeted changes to workflows, and tight control over how data is used. The bank describes its internal ‘LLM Suite’ as a managed setup that allows staff to draft and summarize content using an extensive language model.
Wells Fargo: Output increases ahead of staffing changes
Wells Fargo CEO Charlie Scharf said the bank so far hasn’t cut jobs because of AI, but it is “getting more done.” He said management expects to find areas where fewer people are needed as productivity improves.
In comments reported the same day, Schaaf said the bank’s internal budget already suggests a reduction in its workforce by 2026, even before fully factoring in the impact of AI. He also pointed to the soaring cost of retirement benefits, suggesting preparations are underway for future adjustments.
PNC: AI speeds up long shifts
PNC CEO Bill Demchak positioned AI as an accelerator rather than a new direction. He said the bank’s headcount has remained roughly flat for about a decade, even as its business has expanded. He said that stability comes from automation and branch optimization, and AI could further advance this trend.
Citigroup: Improving software and customer support
Citi’s incoming chief financial officer, Gonzalo Lucchetti, said the bank recorded a 9% productivity increase in software development. This reflects a broader pattern across large enterprises that are deploying AI co-pilots to support coding efforts.
He also pointed to two areas of customer service where AI can help. One is to improve self-service and reduce the number of calls agents receive, and the other is to support agents in real-time when customers need to speak to a human.
Goldman Sachs: Workflow changes combined with hiring restraints
According to Reuters, Goldman Sachs’ internal OneGS 3.0 program is focused on leveraging AI to improve sales processes and customer onboarding. It also targets process-focused functions such as lending workflows, regulatory reporting, and vendor management.
These changes are occurring alongside attrition and a slowing pace of hiring, with workflow redesigns directly tied to staffing decisions.
Where Wall Street banks are seeing the productivity gains from AI the fastest
In any bank, the most obvious results are seen in operations that rely heavily on documentation, follow repeatable procedures, and operate within defined rules. Generative AI can reduce the time needed to find information, summarize material, draft content, and move work through approval chains, especially when combined with structured processes and human checks.
Common areas where we are seeing early impacts include:
Operations: Draft responses, summarize cases, and resolve exceptions faster Software development: Generate code, write tests, refactor, and create documentation Customer service: Powerful self-service sales support and onboarding combined with real-time support for agents: Get data from documents, fill out forms, and set up clients faster Regulatory reporting: Assemble explanations and evidence faster under rigorous review and control
Why governance drives the pace of adoption
For banks, enthusiasm is not the main constraint. The control is. U.S. regulators have long called for strong oversight of models, and that expectation extends to AI systems as well. Guidance such as the Federal Reserve and OCC’s SR 11-7 set standards for model development, validation, and ongoing review. The U.S. Government Accountability Office’s 2025 report states that existing model risk management principles, such as testing and independent oversight, are already being applied to AI.
In practice, this will push banks towards designs that are testable and traceable. The use of AI is often limited by its ability to operate independently. Prompts and output are logged, performance variations are monitored, and humans remain responsible for high-impact decisions such as lending, dispute resolution, and public reporting.
Productivity improves, but employment issues remain
Comments from bank executives suggest a gradual transition. The first phase looks like a combination of stable headcount and higher output as AI tools become more widespread across the team. The second stage begins when benefits are significant enough to impact staffing plans, such as through attrition, role changes, or targeted reductions.
Signals from Wells Fargo about its 2026 headcount plan and severance plans suggest some banks are nearing that second stage.
At a broader level, institutions such as the International Monetary Fund have warned that AI could impact a large proportion of jobs around the world, with different combinations of automation and expansion depending on role and region. The World Economic Forum’s 2025 Future of Jobs Report also predicts significant job mobility as companies adopt AI and adjust their skills needs.
What AI means for Wall Street banking strategy in 2025 and beyond
Banks that gain the most from AI are likely to focus on three areas simultaneously: redesigning workflows rather than layering chat tools, building strong data foundations, and implementing governance that supports speed without compromising trust.
Research firms say the financial risks are high. McKinsey estimates that generative AI could add $200 billion to $340 billion in annual value to the banking sector, primarily through productivity gains.
The open question is no longer whether AI can deliver results in banking. The key is how quickly banks can make these benefits routine, and how they can manage subsequent workforce changes, while maintaining audit trails, security, and customer protection.
(Photo provided by Ro-Ro)
Reference: BNP Paribas introduces AI tools for investment banking
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expos in Amsterdam, California, and London. This comprehensive event is part of TechEx and co-located with other major technology events. Click here for more information.
AI News is brought to you by TechForge Media. Learn about other upcoming enterprise technology events and webinars.
