How to Save Money Now with ChatGPT Finances

Bloch recommends connecting financial accounts and email so ChatGPT can review subscriptions, spot duplicate charges or price increases, and help contact merchants. He demonstrates a 90-day review prompt and says scheduled checks can flag subscription changes each month. He describes the account connection as read-only and says raw transaction and balance data is not used for training.

The dashboard shows spending by category, net worth, investment movements, upcoming cash flow, subscriptions, and bills. Bloch says users can correct transaction categories through the interface or chat, and receive a weekly finance update by email or push notification. Users can also set alerts, such as a warning when monthly DoorDash spending approaches $1,500.

Bloch shows how to ask ChatGPT for tax-saving ideas, estimate possible savings, visualize recommendations, and turn them into a to-do list. He also suggests comparing ETFs and mutual funds for lower-fee alternatives, and reviewing credit card use against rewards and benefits. The finance account connection is available to U.S. users, while people elsewhere can still use connected documents, email, and web search.

Yang describes building interactive sites to explore taking six months off and buying a house, with adjustable assumptions and savings scenarios. Bloch says similar models can help with retirement planning and other decisions, and that users can provide information manually or upload statements when accounts are not connected. The examples are intended to help people compare options and discuss decisions with family.

Bloch describes recurring tasks for subscription checks, household spending reports, and monitoring idle cash. He says financial institutions may block agents from taking investment actions, so execution can vary. Later, he discusses a finance team that shipped 47 changes in seven days and a workflow where engineers use prototypes and AI agents to move from high-level discussions to implementation.