Bank of New York Mellon is doing something that would have sounded absurd five years ago. It’s assigning human managers to oversee teams of artificial intelligence agents — digital workers that will operate alongside flesh-and-blood employees, with their own performance metrics, their own tasks, and their own place on the org chart.
Not as a pilot. Not as an experiment. As strategy.
CEO Robin Vince laid out the vision in stark terms during the company’s investor day in March, describing a future where every manager at BNY will lead a team composed of both humans and AI agents. The bank, America’s oldest financial institution with $52.1 trillion in assets under custody and administration, isn’t tiptoeing into the AI era. It’s sprinting. And the implications — for BNY’s 52,000 employees, for Wall Street’s competitive dynamics, and for the broader question of what work looks like at a major financial institution — are enormous.
“Every manager is going to be a manager of humans and agents,” Vince told investors, as reported by Business Insider. “That is the way to think about what we’re doing.”
Think about that sentence for a moment. The CEO of a 240-year-old bank is telling Wall Street that the fundamental unit of management — the team — is being redefined. Not in some distant future. Now.
From Chatbots to Colleagues: The Rise of Agentic AI in Banking
The concept BNY is embracing goes by the name “agentic AI,” a term that has rapidly migrated from academic papers to boardroom presentations across the financial sector. Unlike traditional AI tools that respond to prompts and wait for instructions, agentic AI systems can independently pursue goals, make decisions within defined parameters, and execute multi-step tasks without constant human oversight. They don’t just answer questions. They do things.
BNY has already deployed more than a dozen AI agents internally, according to Business Insider’s reporting, handling functions that range from coding to client service operations. The bank’s leadership expects that number to grow dramatically. Vince described AI agents as “digital employees” — language that is deliberate and telling. It signals that BNY views these systems not as tools to be used but as workers to be managed.
This isn’t unique to BNY. JPMorgan Chase has invested billions in AI and data infrastructure, with CEO Jamie Dimon repeatedly calling artificial intelligence a defining technology for the firm’s future. Morgan Stanley rolled out an AI assistant powered by OpenAI’s GPT-4 for its wealth management advisors back in 2023. Goldman Sachs has been deploying AI across its trading and compliance operations. But BNY’s framing — putting AI agents on the org chart, giving them human supervisors — represents something more architecturally ambitious.
The bank is essentially proposing a hybrid workforce model where the ratio of humans to AI agents on any given team becomes a management variable, adjusted based on the nature of the work. Some teams might be heavily human. Others might be mostly digital. The manager’s job is to orchestrate both.
And that changes what it means to be a manager at a bank.
Traditionally, managing a team at a financial institution meant hiring, coaching, evaluating performance, handling interpersonal dynamics, and ensuring compliance. Managing AI agents adds a different set of responsibilities: monitoring outputs for accuracy, ensuring agents operate within regulatory guardrails, understanding the logic behind automated decisions, and knowing when to override. It’s a fundamentally different skill set. Or rather, it’s an additional one layered on top of everything else.
BNY has signaled that it’s investing in training its managers for this transition, though specifics remain sparse. The bank has been working with major technology partners including Google Cloud and has been building internal AI infrastructure for several years. Its platform, known internally as Eliza, serves as the backbone for much of the firm’s AI deployment.
The timing isn’t accidental. BNY has been under pressure to demonstrate that its technology investments translate into efficiency gains and revenue growth. The custodian bank model — holding and servicing assets for institutional clients — is a scale business with thin margins. Every basis point of operational efficiency matters. AI agents that can process transactions, reconcile data, generate reports, and handle routine client inquiries faster and more cheaply than human workers represent a direct path to improved margins.
Vince has framed the AI push as essential to BNY’s competitive positioning. During the investor day presentation, he described a company that would be “dramatically more productive” within the next few years, with AI agents handling an increasing share of operational workload. The bank reported $1.2 billion in technology spending in 2024, a figure that’s expected to grow.
The Workforce Question Nobody Wants to Answer Directly
Here’s the tension at the heart of BNY’s strategy: if AI agents are digital employees, what happens to the human ones?
Bank executives across the industry have been careful to frame AI as augmenting human workers rather than replacing them. Vince has echoed this line, suggesting that AI agents will free up employees to focus on higher-value tasks — client relationships, complex problem-solving, strategic thinking. It’s a familiar refrain. And it’s not entirely wrong. But it’s incomplete.
The math is straightforward. If an AI agent can do the work of two or three people in certain operational functions, and BNY deploys hundreds or eventually thousands of these agents, the headcount implications are real. The bank doesn’t need to announce layoffs to reduce its workforce. Attrition, hiring freezes, and reorganization can accomplish the same thing gradually and with far less public scrutiny.
Business Insider’s reporting noted that BNY’s leadership views the AI agent deployment as a way to handle growing volumes of work without proportionally growing headcount. That’s a polite way of saying productivity per employee goes up — which, in a business with relatively flat revenue growth, eventually means fewer employees are needed to generate the same output.
This dynamic isn’t unique to BNY. Citigroup announced in early 2025 that it would cut 20,000 jobs as part of a broader restructuring, with technology and automation playing a role. Deutsche Bank has been trimming its workforce for years while increasing technology investment. The pattern is consistent: banks spend more on technology, they need fewer people for back-office and middle-office functions, and the workforce contracts — even as executives insist that AI is creating new roles.
Some new roles will indeed emerge. AI governance specialists. Prompt engineers. Agent supervisors. But the volume of new positions is unlikely to match the volume of displaced ones. Not even close.
For BNY’s 52,000 employees, the message from leadership is clear even if it’s never stated explicitly: adapt or risk obsolescence. The managers who thrive will be those who can effectively orchestrate hybrid human-AI teams. The individual contributors who thrive will be those whose skills complement rather than compete with what AI agents can do.
So what can’t AI agents do? For now, quite a lot. They struggle with ambiguity. They can’t build trust with a nervous client over lunch. They don’t understand institutional politics or read the room in a board meeting. They can’t exercise the kind of judgment that comes from decades of experience in a specific market. But the list of things they can’t do is shrinking, and quickly.
The regulatory dimension adds another layer of complexity. Financial regulators in the U.S. and Europe have been increasingly focused on AI governance in banking. The Office of the Comptroller of the Currency, the Federal Reserve, and the SEC have all signaled that they expect banks to maintain human oversight of AI-driven decisions, particularly in areas that affect consumers and market integrity. BNY’s model of assigning human managers to AI agent teams may, in part, be a preemptive response to regulatory expectations — ensuring there’s always a human in the loop, at least nominally.
But nominal oversight and effective oversight are different things. If a manager is supervising fifteen AI agents processing thousands of transactions per hour, the practical ability to meaningfully review each decision is limited. The manager becomes less of a supervisor and more of an exception handler — stepping in only when something goes wrong. That’s a model that works until it doesn’t.
The risks are real. AI agents can hallucinate — generating plausible but incorrect outputs. They can amplify biases embedded in training data. They can make errors that compound at machine speed. In a custodian bank handling trillions of dollars in assets, even small errors can cascade into significant problems. BNY’s internal controls and risk management frameworks will need to evolve as fast as its AI deployment — a challenge that’s easier to describe than to execute.
What This Means for the Rest of Wall Street
BNY’s move puts pressure on every other major financial institution to articulate its own AI workforce strategy. It’s one thing to talk about AI in vague, aspirational terms. It’s another to say, as Vince did, that every manager will lead a mixed team of humans and machines. That’s a concrete organizational commitment, and it forces competitors to respond.
The competitive dynamics are particularly acute in the custody and asset servicing space, where BNY competes directly with State Street and JPMorgan’s securities services division. If BNY can deliver the same services with meaningfully lower costs because AI agents are handling a larger share of operational work, competitors will have to match that efficiency or lose business. It becomes an arms race — not for the best talent, but for the best combination of talent and technology.
And the technology providers are paying attention. Microsoft, Google, Amazon, and a growing roster of AI startups are all competing to be the infrastructure layer beneath these enterprise AI deployments. Google Cloud’s partnership with BNY is a significant reference case. If BNY’s hybrid workforce model delivers results, it becomes a template that Google can sell to other financial institutions. The same dynamic plays out with Microsoft’s Azure OpenAI Service, which powers AI deployments at several major banks.
The consulting firms are circling too. McKinsey, Deloitte, Accenture, and others have all published extensive research on agentic AI in financial services, and they’re actively pitching transformation engagements to bank leadership teams. The market for AI strategy consulting in banking is booming — which tells you something about how much uncertainty exists at the executive level about how to actually implement these changes.
For BNY’s investors, the key question is execution. The vision Vince articulated is compelling on paper. But large-scale organizational transformation at a 240-year-old institution with complex legacy technology systems, entrenched processes, and tens of thousands of employees is extraordinarily difficult. Many ambitious technology transformations in banking have underdelivered. The history of enterprise software implementations at major banks is littered with cost overruns, delays, and disappointing results.
BNY’s stock has performed well in recent years, partly on the strength of its technology narrative. The market is giving the bank credit for its AI ambitions. But that credit comes with expectations. If the productivity gains don’t materialize — if the AI agents prove less capable than advertised, if regulatory pushback slows deployment, if the cultural change required to make hybrid teams work proves too difficult — the stock will reflect that disappointment.
What’s not in doubt is the direction. The question of whether AI agents will become a standard part of the banking workforce has been answered. They will. BNY is simply saying it out loud, putting structure around it, and moving faster than most. Whether that speed translates into sustainable competitive advantage — or just expensive early-mover lessons — is the story that will unfold over the next several years.
One thing is certain: the org chart at America’s oldest bank will never look the same again.
WebProNews is an iEntry Publication