Major banks are stepping up efforts to embed digital assistants into everyday workflows, refining how those agents interact with staff and clients as institutions compete to capture productivity gains from agentic AI.
Agentic AI refers to systems that can complete tasks with minimal human input. Financial firms are now exploring versions of those systems that can autonomously perform actions on users' behalf while operating alongside human employees. Use cases span wealth management, client onboarding and vetting, trading operations and corporate treasury work.
“We are working with banks in particular on agents and human employees ... to help the banks look at all the roles end to end, and then determine which ones are hybrid roles, which ones are agentic employees, which ones are only human employees,” said Peter Torrente, U.S. sector leader for banking at KPMG. A June KPMG survey cited by Torrente found that 51% of banks were piloting AI agents.
Executives at several large banks described practical deployments and near-term tests of digital assistants. Koren Maranca, head of Artificial Intelligence for Wealth Management at Morgan Stanley, said the firm plans to begin testing client-facing digital assistants later this summer. Morgan Stanley already employs agents to help financial advisors with a variety of tasks.
“We are now preparing these agents to start pushing reminders or recommendations to the financial advisors regarding their clients,” Maranca said. Within wealth management, these assistants can analyze investments, offer strategy suggestions and help construct portfolios.
At BNY Mellon, management has framed digital employees as teammates assigned to specific tasks that can communicate with one another. CEO Robin Vince described on a Wall Street Journal podcast how the bank has given some digital assistants login IDs and nicknames so they can join teams and coexist with human colleagues.
“The digital employee has a login, it can actually operate in the systems, and it actually has a ... human manager that’s responsible for training it, making sure that it actually is doing all the right things, like a performance review, if you will, quality control, and it has tasks every day,” Vince said on the podcast while explaining the role of a digital assistant named Payment Pete. BNY did not respond to requests for comment for this article.
UBS has also integrated agents into advisor workflows. Richard James, head of AI product at UBS, said the bank’s systems collect internal information from meetings, accounts and e-mail communications to generate alerts when action is required - for example, when an annuity is maturing and reinvestment is needed.
“They gather all internal information from meetings, accounts and e-mail communications,” James said. UBS reports that once an advisor decides to proceed with a client transaction, agents can execute trades and complete money transfers. The bank says AI is enabling advisors to spend roughly 70% of their time speaking with clients rather than performing routine tasks.
Other banks are pursuing partnerships and pilots to expand agentic capabilities. Goldman Sachs formed a collaboration earlier this year with Anthropic to build agents that will support tasks such as trading and transaction accounting, client vetting and onboarding. JPMorgan has identified areas like corporate treasury as particularly susceptible to transformation by agentic AI, and Citi is preparing to introduce an AI-enabled virtual wealth management “team member.”
“Banks are increasingly using agentic AI and figuring out more ways to use it because it has a lot of potential,” said Bhavi Mehta, global lead for advanced analytics in financial services at Bain & Company.
With technology spending on the rise, investors have pressed banks on expected returns. Torrente noted that investor questions about return on investment are influencing where banks concentrate AI spending.
“Investors are asking, where should we be looking for ROI on these tech spends and that’s why banks are likely to focus on certain areas of AI spends where the returns are much more evident and can be scaled up,” Torrente said.
Accountability and oversight concerns
The broader rollout of agentic AI is prompting executives and regulators to scrutinize the risks attendant to giving digital agents access to internal systems. Firms say they are implementing guardrails and supervision while testing ways the technology can safely interact with customers and internal workflows.
Morgan Stanley’s leadership has emphasized that human oversight will remain in place and that agents will not be granted autonomy to make portfolio decisions. Mehta reinforced that banks are exercising caution when agentic systems touch customers.
“They are still primarily using it for internal purposes and are being extremely cautious when it touches the customer and are making sure that there is a human involved for any critical functions,” Mehta said.
As banks move from pilots to broader deployments, the industry faces a balancing act - scaling agents to boost efficiency while tightening controls, defining human-versus-agent responsibilities, and satisfying investor demands for measurable returns.