Anthropic announced Personal Finance in Claude on November 5, a beta feature that lets paying subscribers in the United States connect their bank and card accounts to the assistant and ask about their money in plain language. The connection runs through Plaid, access is read-only, and Claude cannot move, manage or invest your money. The rollout is gradual, reaching Pro, Max and Ultra subscribers first, with no date yet for other plans or countries.
What Claude can see — and what it cannot
For each account you choose to connect, Claude sees the bank’s name, the account’s name, type and last four digits, its balance and credit limit, and up to two years of transactions. You sign in to your bank yourself through Plaid, so Claude never asks for or sees your password.
You can unlink an account at any time, and Claude then loses access to its data. Conversations inside the Personal Finance hub are not used to train Anthropic’s models, and the feature carries no ads or sponsored recommendations. Anthropic also offers no credit cards, loans or savings accounts of its own.
Anthropic’s worked example shows what the answers look like. Asked where the money went, Claude adds up spending and sorts it by category, reaching $4,309.37, with more than half going to rent and utilities and $1,024.51 to food and drink. Follow-up questions surface a membership renewal, four payments due before payday, and $1,393.22 left over.
The limits Anthropic puts in writing
Personal Finance is not designed to give investment advice. Anthropic is unusually direct about the accuracy risk: it warns that Claude can make mistakes, and tells users to ask which transactions sit behind a number and check important figures with their bank. Its Help Center adds that balances can lag, and that an account linked twice can be counted twice.
The guest list is narrow on purpose. The beta covers Pro, Max and Ultra plans in the United States only, leaving out Free, Team and Enterprise subscribers, and the section does not appear outside the US at all. Those are the same $100 and $200 tiers a federal judge ruled may have been marketed with misleading usage promises.
Our Read
The read-only boundary is the whole design. Every rival assistant is racing to touch personal data, and the first version that moves money by mistake would set the category back years. Anthropic chose the cautious path: answers first, no transactions, no ads, no house financial products.
That restraint is the feature, not a limitation to be fixed later. Claude is late to this race, and that may help it: ChatGPT’s finance features arrived in May, according to coverage of that launch, so Anthropic had six months to watch what users actually asked for.
What Anthropic built is narrower and more careful, and it lands on a subscriber base that already pays premium subscription prices — Claude already takes 29% of UK consumer AI subscription spending, as our report on its UK surge shows.
The risk is not the plumbing but the answers. Two years of transactions is the most intimate data an assistant can hold, and a wrong total could mislead someone about money they do not have. Anthropic’s caveat, printed in the announcement itself, is the right tone — but the test will be how often users check the numbers behind them.
Why It Matters
Personal finance is the most sensitive data a general assistant can touch, and whoever earns trust here sets the terms for everything that follows. Anthropic says it wants Claude to work alongside licensed professionals and regulated firms on bigger money decisions.
The company is inviting banks and wealth managers to help shape that. If the beta holds up, the assistant moves from answering questions about your money to advising on it — a bigger prize, and a bigger risk, since advice carries liability that answers do not.
Background
Claude’s 2026 has been a steady expansion from chatbot to workplace tool: message threads for long-running agents arrived in public beta on November 2, and Anthropic filed to go public in October. Personal Finance is the consumer-facing step, putting the assistant inside daily life rather than just daily work.
Source: Anthropic’s announcement and Unite.AI’s report
0 Comments