Morgan Stanley is preparing to let outside artificial intelligence agents connect directly to parts of its wealth management infrastructure, a step that could reshape how corporate clients manage stock plans and related services. The bank is among the first major Wall Street firms known to be opening internal systems to external AI tools rather than keeping those tools confined within the company.
The planned rollout will apply to Morgan Stanley at Work, the unit that handles stock administration platforms including ShareWorks and Equity Edge. According to Mark Mitchell, the business’s chief product officer, clients’ AI agents will be able to retrieve data and insights from those systems without relying on the traditional software interface designed for human users.
Mitchell said the bank envisions a future in which corporate clients are no longer logging into those platforms themselves. Instead, he said, they will use agentic AI tools inside their companies to interact with Morgan Stanley’s systems on their behalf.
The bank has already provided a small group of clients with early access, and it plans to expand that access to all of its 3,400 administration clients by next year, Mitchell said. CNBC reported the plans exclusively.
The move reflects a broader shift on Wall Street as financial firms explore what AI agents can do in place of, or alongside, human users. Banks including JPMorgan Chase and Goldman Sachs are already using AI agents internally for tasks such as code generation. But neither has publicly announced plans to let outside agents connect directly to their internal systems.
Morgan Stanley’s strategy is closely tied to the role workplace stock plans play in its broader wealth management business. The firm has turned stock compensation administration into a pipeline for its advisory division, which is now the world’s largest wealth management business with $7.35 trillion in client assets.
By handling employee stock plans for corporations, the bank can build relationships with workers as their personal wealth grows. Morgan Stanley has said that approach has helped it gather $1.2 trillion in workplace-related assets. The bank expanded that business with its acquisitions of Solium Capital in 2019 and E-Trade in 2020.
The company says the workplace business now serves a large share of major corporations, including nearly half of the companies in the S&P 500 and eight of the 10 largest unicorn startups. Mitchell said the appeal of AI for corporate clients is straightforward: fast-growing technology and biotech firms want to manage increasingly complicated stock plans without hiring additional support staff.
Morgan Stanley also sees an internal efficiency gain. Mitchell said agentic AI could help the firm scale customer support and plan administration without adding thousands of employees. For the bank, the change is less about whether users continue to visit its websites and more about keeping control of the underlying data and business logic.
To make the system work, Morgan Stanley is relying on the Model Context Protocol, an open-source standard that allows AI models to connect to data sources. The bank, which began working with OpenAI in 2022, believes the rise of AI agents will change how enterprise software is accessed and used.
Mitchell said the company is not concerned that clients may bypass the web-based front door to its platforms. In his view, firms that pair proprietary data with proprietary business logic will remain valuable even if the user interface shifts away from traditional logins.