Major U.S. banks are moving beyond experimental chatbots and integrating autonomous AI agents into core operations, including trading, treasury, and wealth management. These digital employees now handle tasks such as client vetting, trade execution, and proactive advisor alerts, according to a Reuters report published Tuesday. The shift reflects a broader industry race to determine the optimal balance between human and machine roles.
Peter Torrente, U.S. banking sector leader at KPMG, told Reuters that his firm is working with banks to evaluate roles end-to-end, deciding which should be hybrid, fully automated, or exclusively human. A KPMG survey last month found that 51% of banks are already piloting AI agents, signaling rapid adoption across the sector.
Wealth Management Gets a Digital Assistant
At Morgan Stanley, the focus is on advisor support. Koren Maranca, head of AI for wealth management, said the firm will begin testing digital assistants later this summer that can interact with clients around the clock. Existing agents already handle routine tasks, and the next phase involves agents that proactively push reminders and recommendations to advisors regarding client needs. Separately, Morgan Stanley announced in early June that it would open its stock-plan administration platforms, ShareWorks and Equity Edge, to external AI agents built by corporate clients. Mark Mitchell, chief product officer of Morgan Stanley at Work, said the firm expects corporate clients to eventually stop logging into these systems, instead using agentic tools that pull data via the Model Context Protocol. The unit oversees roughly $1.2 trillion in assets and serves nearly half the S&P 500, with plans to extend agent access to all 3,400 stock-plan clients by next year.
At UBS, AI agents generate thousands of daily alerts for advisors, flagging events such as an annuity nearing maturity that requires reinvestment. Richard James, head of AI product at UBS, explained that agents gather internal information from meetings, accounts, and email communications. Once an advisor approves a transaction, the agents can execute trades and money transfers.
Digital Employees with Logins and Performance Reviews
BNY has taken the concept further by treating digital employees as teammates with their own login credentials and nicknames. CEO Robin Vince described one such assistant, nicknamed Payment Pete, on a Wall Street Journal podcast earlier this year. The digital employee has a login, operates within systems, and is assigned a human manager responsible for training, quality control, and performance reviews.
JPMorgan has advanced on the investing side. A July 9 research note detailed eight AI-powered agents built on models from OpenAI and Anthropic that classify markets into four regimes—Goldilocks, reflation, stagflation, and risk-off—and adjust allocations accordingly. In backtests spanning roughly two decades, every agent outperformed a traditional 60/40 stock-bond portfolio on a risk-adjusted basis, with the best performer adding 0.7 percentage points of annualized return at lower volatility.
Anthropic rolled out a suite of financial-services agents in early May covering pitchbooks, credit memos, KYC screening, and statement audits, appearing alongside JPMorgan CEO Jamie Dimon at a New York event. Goldman Sachs partnered with Anthropic earlier this year to automate trading, transaction accounting, and client onboarding. JPMorgan also sees corporate treasury as ripe for an agentic overhaul.
Investors Seek ROI, Regulators Demand Guardrails
Bhavi Mehta, global lead for advanced analytics in financial services at Bain & Company, noted that banks are increasingly finding ways to use agentic AI due to its potential. However, investors are pressing for clear returns on technology spending. Torrente said banks are likely to focus AI investments on areas where returns are evident and scalable.
The speed of adoption has drawn regulatory scrutiny, given that agents are being granted access to internal systems even as oversight frameworks are still being built. Maranca emphasized that Morgan Stanley will maintain human oversight at all times and that agents will not have autonomy over portfolio decisions.
For advisors, the rise of agentic AI parallels trends in the broader wealth management industry. A recent HSBC survey found that 57% of affluent investors use AI for research, but 59% still trust human advisors for final decisions. Meanwhile, Salesforce has deployed its own agentic AI suite to counter third-party notetaker threats in wealth management, and Zocks debuted an agentic AI tool for advisors to scan books for growth gaps.
