Story
Nov 5, 2026

Claude Wants to Read Your Bank Statements, Not Move Your Money

Anthropic is pitching Claude as a neutral interpreter of everyday money questions, while the obvious trade-off is trust: users gain a detailed view of their spending only by granting an AI access to some of their most sensitive financial data.

Anthropic on Thursday began rolling out Personal Finance, a beta feature that lets U.S. subscribers connect bank and credit-card accounts to Claude through Plaid. The company’s pitch is simple: ask the chatbot where the month’s money went, what bills remain before payday, or whether a major purchase is affordable, and receive an answer in ordinary language rather than a spreadsheet.

The tool arrives after OpenAI introduced a comparable Plaid-connected finance feature for ChatGPT in May, putting Claude in a race to become a more useful — and more intimate — financial assistant. Anthropic’s approach is deliberately framed as analysis, not control: Claude cannot transfer funds, manage accounts or make investments, and is not designed to provide investment advice.

In practice, Claude can sort spending into categories such as rent, food and shopping; compare current habits with previous periods; and flag duplicate charges, surprise fees or bills that are creeping upward. It can also explain concepts such as APR and calculate interest paid over the year. Anthropic says it offers “answers in everyday language” from a user’s balances and transactions, without ads, sponsored recommendations or its own financial products.

That restraint does not erase the privacy question. To perform the analysis, Claude may see a bank’s name, account names and types, the last four digits of account numbers, balances, credit limits and as much as two years of transaction history. Plaid handles the connection, meaning Claude does not receive banking passwords; Anthropic also says Personal Finance conversations are not used to train its models.

For now, the feature is available to Pro, Max and Ultra subscribers in the U.S. Anthropic calls it a first step, arguing that bigger decisions should involve licensed professionals and regulated firms — an acknowledgment that a chatbot may clarify the numbers, but cannot replace the people responsible for acting on them.