AI Quietly Replaces White-Collar Tasks as Banks Embrace Digital Employees
BNY Mellon and JPMorgan Chase lead a shift toward AI systems that mirror white-collar labor, raising new questions about the future of office work.
Financial giants BNY Mellon and JPMorgan Chase are introducing artificial intelligence systems that replicate human workflows, a move that reflects a broader shift in white-collar employment structures, according to a report by The Wall Street Journal.
When AI Gets a Desk and a Manager
At BNY Mellon, AI-driven “digital employees” have been assigned company logins, managers, and narrowly scoped responsibilities. These agents now operate across software engineering and operations teams, handling tasks like detecting security flaws in code and validating payment instructions—work historically managed by salaried staff.
Leigh-Ann Russell, the firm’s Chief Information Officer and Global Head of Engineering, said the digital workers aren’t just tools but modeled deliberately after human colleagues, with individual assignments and access permissions. “This is the next level,” she said, anticipating broader rollout in the coming months.
The bank expects to expand these agents' roles further by granting access to internal communications platforms such as email and Microsoft Teams, allowing them to reach out to human coworkers when they hit operational limits.
Redrawing the White-Collar Line
The deployment signals a subtle but growing shift from automation as support infrastructure to automation as peer labor. Each AI persona created by BNY’s internal hub is duplicated in dozens of instances across small teams, with careful access controls to limit their visibility across the organization.
Despite those limits, the systems operate with a surprising degree of independence. An AI engineer can identify a vulnerability, patch the flaw, and submit it for approval without human involvement until the final step.
This division of labor suggests that white-collar work is being unbundled—analyzed, dissected, and reassigned—so that machines can handle high-frequency, high-precision tasks that previously required human judgment.
The Quiet Replacement Pipeline
JPMorgan Chase has taken a parallel path. While it hasn’t yet granted its AI systems full login access, the bank has equipped some 200,000 employees with a generative AI platform called LLM Suite. The software helps staff write summaries, brainstorm ideas, and automate documentation—functions once tied to assistants, analysts, or entry-level roles.
Chief Analytics Officer Derek Waldron said the term “digital employee” is meant to help teams conceptualize how these tools fit within organizational structures. The tools are “neither just human nor purely software,” he noted, and require new systems of access, compliance, and accountability.
JPMorgan is also exploring tailored deployments, where job-specific versions of the chatbot evolve into autonomous role players. Waldron emphasized that these changes are not one-size-fits-all. “It has to be figured out on a case-by-case basis,” he said.
A Blurred Future for Human Roles
Across the industry, questions remain about what happens when digital agents start working side by side with people—particularly when those agents operate faster, don’t take breaks, and don’t demand promotions.
Scott Mullins, Managing Director for AWS Financial Services, said integration is the central challenge. “How do we coordinate that work together? How do we manage those folks? What’s the new operating model?” he asked.
Unlike legacy automation tools that lived behind the scenes, these systems are now visible in workflows—proactively flagging anomalies, writing updates, and soon, potentially chatting with colleagues. That shift raises not only logistical questions but also existential ones about the meaning of work and the permanence of human job functions.
Human Talent Still in the Loop—for Now
Both institutions stress that digital employees won’t replace hiring pipelines anytime soon. BNY Mellon continues to recruit engineering talent, and JPMorgan is expanding its AI ethics and oversight teams.
But both are also pushing AI deeper into core business functions. Tasks once thought immune to machine replication—code debugging, financial validation, compliance flagging—are now being partially handed off. And with each successful deployment, the boundary between support tool and substitute worker grows harder to define.
The clearest sign of what’s changing may be found not in any single system, but in the language: banks are no longer talking about AI tools. They’re talking about employees.