The feature, announced November 5, is rolling out in beta to US subscribers on Claude’s Pro, Max and Ultra plans. Users can connect selected financial accounts and ask questions about spending, recurring bills, savings and cash flow without manually consolidating information from multiple banking applications.
What Financial Information Can Claude Access?
Claude can access the names and types of accounts users authorize, their balances, credit limits, the last four digits of account numbers and up to 24 months of transaction history. For investment and loan accounts, the current integration does not provide detailed holdings, trading activity, interest rates or contractual terms.
Users authenticate through Plaid rather than providing banking credentials directly to Anthropic. They can also disconnect accounts, although financial information already included in previous conversations may remain in their chat history until deleted.
The access is strictly read-only. Claude cannot initiate transfers, place investment orders or independently manage customer funds. Anthropic also said the feature is not designed to provide investment advice.
The company has excluded advertisements and sponsored financial recommendations from the product. Conversations within the dedicated Personal Finance hub will not be used to train Anthropic’s models, although information shared outside that environment remains subject to the user’s applicable privacy settings.
These restrictions separate account-level financial analysis from transaction execution. Consumers can ask how much they have spent or whether their cash flow supports a planned purchase, but the platform cannot act on those conclusions by transferring or investing funds.
Investor Takeaway
Read-only connectivity gives AI platforms access to customer financial data without granting control over assets, limiting execution risk while leaving accuracy and privacy concerns.
Why Does the Plaid Integration Matter for Banks?
Plaid provides the infrastructure connecting consumer financial accounts to third-party applications. Its role in Claude’s launch places established open-banking networks between financial institutions and AI platforms seeking to offer personalized services.
That arrangement raises commercial questions about the ownership, distribution and pricing of financial data. Banks maintain the underlying customer accounts, but AI platforms increasingly control the interface through which consumers interpret their finances.
The economics of these connections are already changing. JPMorgan reached a data-access agreement with Plaid in September 2025 that introduced fees for access to customer account information. The arrangement demonstrated that consumer-authorized financial connectivity can involve substantial commercial negotiations between banks and data aggregators.
As account-connected AI services expand, banks and data networks will need to address authentication, consent management, data-access costs and the handling of potentially sensitive financial information.
How Is Charles Schwab Handling External AI Access?
Charles Schwab has taken a more controlled approach to autonomous AI agents accessing customer accounts. Chief Executive Rick Wurster said the brokerage blocks unauthorized agents attempting to log in on behalf of clients, citing authentication and security concerns.
Schwab has also introduced Charley, its own AI-powered assistant, which draws on the firm’s internal information and security controls to provide customer support and financial insights.
However, Schwab is not excluding external AI technology altogether. The brokerage has partnered with Anthropic to make Claude available to more than 16,000 independent investment advisers served by Schwab Advisor Services.
Schwab is also developing controlled connections between customer-selected AI models and its own infrastructure, allowing the brokerage to manage authentication and permissions rather than permitting unrestricted automated access.
The distinction is important. Anthropic’s Plaid integration relies on consumer-approved data sharing, while Schwab’s approach emphasizes institution-controlled connections and restrictions on autonomous actions. Both models allow AI to analyze financial information, but they allocate control over access differently.
Investor Takeaway
Banks may accept AI-powered customer services while retaining control over authentication, permissions and transactions, making approved data connections commercially important.
What Comes Next for AI-Powered Financial Services?
Anthropic has described Personal Finance as an initial offering and said it is exploring ways for Claude to work alongside licensed financial professionals and regulated institutions.
That distinction could become increasingly important as AI applications progress from explaining historical transactions toward making recommendations based on individual financial circumstances.
There are also practical limitations. Anthropic warns that account information may not always be current and that Claude can make calculation errors. Incomplete connections or duplicated accounts could lead to misleading estimates of available funds.
For banks, brokerages and financial technology providers, the competitive question is no longer simply whether AI can explain financial information. It is which institutions will control access to that information, who will be responsible for the accuracy of financial guidance and how far AI platforms will be permitted to act on behalf of customers.
