[Summary]

On May 15, 2026, OpenAI announced Finances, a new personal finance experience that allows users to connect their financial accounts on ChatGPT and view spending, net worth, investment allocation, subscriptions, payment schedules, and more.

The service will first be available to ChatGPT Pro users in the United States, and will be rolled out gradually through the web version and iOS app. Plaid will be used to connect with financial institutions, and OpenAI explains that it supports over 12,000 financial institutions.

The essence of this announcement is not the addition of household accounting functions. General-purpose AI is moving to take control of financial information that has been scattered across banking apps, securities apps, credit card statements, and household account book apps. On the other hand, financial data is information that is closely related to our daily lives, and the line between convenience and privacy is once again being questioned.

First, the conclusion

ChatGPT Finances has the potential to shift the competitive axis of personal financial services from "recording" to "consultation."

Traditional household account book apps focused on expense classification, graphs, and visualization of monthly income and expenditure. Finances aims to take decision support a step above that.

Traditional household budget managementAreas targeted by ChatGPT Finances
Record your expensesConfirm the meaning of your expenses through conversation
View graphDiscuss improvement measures
Switch appsOrganize the situation across accounts
User analyzesAI presents patterns and arguments
See the pastSimulate future options

In other words, OpenAI's target is not just a household accounting app, but the personal finance user experience itself.

What can you do?

OpenAI explains that Finances is a feature that connects financial data such as bank accounts, credit cards, and brokerage accounts to ChatGPT, allowing you to check your financial status in conversations and dashboards.

The main uses are as follows.

AreaWhat you can do
Spending ManagementCheck spending by category, this month's spending, and recent trends
Fixed costsCheck subscriptions, billing schedules, and duplicate payments
Asset managementCheck net worth trends and portfolio allocation
Confirm investmentsUnderstand the holding status of stocks, ETFs, etc., and daily fluctuations
PlanningConsultation on budgets, savings, loan repayments, and plans for major purchases

The important thing is that it's not just a list of numbers.

For example, it is designed to answer questions such as ``Where did my spending increase this month?'' ``Will I stay within my budget at this rate?'' and ``How can I realistically pay off my credit card debt?'' based on connected data.

Financial assistant, not a “household account book”

The change this feature represents is that the focus of household finance management shifts from the dashboard to dialogue.

With the household account book app, users had to look at graphs, find the cause on their own, and think about their next course of action.

Finances allows users to ask questions in natural language.

User questionsValue on the AI side
What is the reason for this month's increase in spending?Organize increase categories from transaction data
Are there any subscriptions that are candidates for cancellation?List recurring charges and provide material for considering the value of use
What to check before purchasing a homeOrganize the points based on income, expenses, savings, and debt
Is your investment allocation biased?Check the distribution and risk of your assets

In this structure, the person the user talks to is more important than the screen the user touches.

If the entrance to financial UX changes from "opening a bank app" to "asking ChatGPT," the position of financial institutions and household accounting apps will change significantly.

What Plaid and Intuit mean

Plaid is used to link Finances accounts. Plaid is a widely used foundation for financial data connectivity in the United States, and OpenAI says it supports more than 12,000 financial institutions, including banks, credit cards, and brokerages.

OpenAI also mentions its collaboration with Intuit.

What we see here is that rather than becoming a financial institution itself, OpenAI is creating a conversational upper layer on top of existing financial data infrastructure and accounting/tax ecosystems.

LayerRole
Financial institutionsOriginal data such as accounts, cards, securities, loans, etc.
Plaid etc.Financial data connection/permission management
Intuit, etc.Practical connections such as taxation, accounting, application, etc.
ChatGPTCross-sectional understanding, consultation, organization, decision-making support

Competition among financial apps is becoming more and more a competition not only in terms of the number of functions, but also in ``which screen and which conversation can start from?''

Positions targeted by OpenAI

What OpenAI is looking for is a consultation point before decision-making rather than the sale of financial products itself.

If ChatGPT becomes the first service users open when they think about buying a home, changing jobs, repaying loans, allocating investments, organizing subscriptions, or checking taxes, financial institutions will lose some of their customer contacts.

This is similar to the changes happening in the search market.

In the past, search engines were the gateway to searching for something. Nowadays, the number of people asking AI directly is increasing. If the same thing were to happen in finance, banks, securities companies, household accountants, robo-advertisers, and credit card companies could lose the point of contact even though they have the final execution function.

Design for safety

OpenAI describes Finances as a design that prevents ChatGPT from directly performing financial actions.

The Help Center lists things that ChatGPT cannot do, such as sending money, paying bills, changing account settings, trading, changing retirement contributions, opening and closing accounts, filing tax returns, and acting as a financial, legal, tax, and investment advisor.

In other words, Finances at the moment is not ``AI that moves assets,'' but ``AI that understands and organizes financial situations.''

This difference is important. By not having execution privileges, the direct risk of erroneous operations and fraudulent remittances is reduced.

Risks that still remain

However, just because it's for viewing only doesn't mean it's safe.

Financial data is more than just numbers.

DataReadability
Expenditure historyLifestyle patterns, hobbies, friendships
Account balanceFinancial capacity, vulnerability
Investment accountRisk tolerance, asset formation policy
LoansHousing, education, cars, borrowing burden
SubscriptionAreas of interest, services used, guessing family structure

If your ChatGPT account is compromised, you may not be able to send money, but there is a risk that your highly accurate personal profile will be leaked.

Additionally, financial counseling is connected to users' concerns and future plans. The more asset information, life goals, family composition, and debt status accumulates in conversations with AI, the more important account protection, data deletion, and usage settings become.

OpenAI provides guidance on deleting synced data after disconnecting, managing financial memories, not accessing connected financial data in temporary chats, and using multi-factor authentication. Users also need to at least enable MFA and regularly review unnecessary connections and memories.

AI-specific risk of prompt injection

When using AI in the financial field, it is necessary to be aware of not only conventional information leaks, but also the attack surface unique to AI.

One of them is prompt injection.

This is an attack that uses external information or malicious input to guide AI decisions and outputs. Even if Finances itself cannot perform remittances or transactions, the risk of malicious information being mixed into decision support increases as it becomes more integrated with external applications and financial services in the future.

For example, fake investment information, fraudulent tax guidance, and overly fear-mongering financial content, when combined with a user's individual financial situation, can be a very convincing fraud vector.

For this reason, AI financial UX requires not only "convenient answers," but also evidence display, operational restrictions, external input verification, and expert confirmation when making important decisions.

Suggestions for the Japanese market

Finances is currently in preview for US Pro users. It is undecided whether it will be offered in the same format in Japan.

However, the impact on the Japanese market cannot be ignored.

In Japan, competition is already underway to expand financial contacts with companies such as Money Forward, Zaim, banking apps, securities apps, Rakuten Economic Zone, SBI Economic Zone, and Sumitomo Mitsui Olive.

When general-purpose AI enters this field, the axis of competition will change.

So farPossible changes
Financial institutions and household account book apps are the gatewayGeneral-purpose AI is the gateway for consultation
View data by appAI organizes data horizontally
Complete within economic areaCompare across economic areas
UI of financial products takes center stageUX of decision support takes center stage

Japan's household account book apps and financial super apps have strengths in data linkage, classification accuracy, and financial institution cooperation.

On the other hand, general-purpose AI such as ChatGPT can cross not only household finances but also life contexts such as work, travel, learning, health, and family planning. The biggest change is that finance will be incorporated into AI for all aspects of life, rather than just financial apps.

Points that investors should look at

This news is not only the announcement of AI functionality for individuals, but also material that will change the competitive structure of the fintech market.

From an investor's perspective, the following points are important.

Issues to look atMeaning
Customer contact points for financial institutionsCan banking and securities apps maintain their entry points
Differentiation of PFM appsRecording and classification alone are easily absorbed by AI
Data connectivity infrastructureAPI infrastructure like Plaid is becoming more important
Security investmentReliability is a competitive advantage in AI that handles financial data
Regulatory complianceBoundary management between investment advice, taxation, and personal information protection is necessary

It is particularly important to note that AI will not immediately replace financial institutions.

Accounts, payments, securities transactions, tax returns, identity verification, and regulatory compliance continue to be the domain of specialized businesses. However, if AI becomes the first point of contact for users, financial institutions may be pushed down to the "execution layer."

Summary

ChatGPT Finances should be seen as the beginning of a race to become the gateway to personal finance UX, rather than a new feature for household accounting apps.

Users enjoy the convenience of conversationally organizing their spending, assets, investments, subscriptions, and payment schedules.

For the financial industry, general-purpose AI is moving to take over customer contact points that were previously held by banks, securities, household account books, robo-ads, and credit card companies.

However, behind the convenience comes the risk of aggregating financial data into AI.

The question from now on is not just how accurately AI can answer questions.

To what extent should we hand over data close to our lives, such as our income, expenses, assets, debts, and future plans, to AI?

Finances is a feature that brings that question to the heart of personal finance.

source

This article is for educational and informational purposes only, based on public information. It is not a recommendation or solicitation to buy or sell any specific security or financial product. Although care is taken with accuracy, the content and future investment outcomes are not guaranteed. Final investment decisions should be made at your own judgment and responsibility.