OpenAI Introduces ChatGPT Wallet for Automated Purchases

OpenAI launched ChatGPT Wallet on November 13, letting its Dots AI agents buy products and book hotels on a user's behalf. Users must approve every purchase with Face ID or a fingerprint scan before the agent can pay. The feature is rolling out to eligible ChatGPT Pro subscribers in the US.

This development marks a shift in how artificial intelligence interacts with everyday digital commerce. Rather than merely recommending products or drafting emails, the software now holds financial utility, capable of completing transactions across the internet. As artificial intelligence integration deepens, companies are exploring ways to reduce friction in online retail by delegating repetitive tasks to software assistants.

The system relies on specialized AI agents known as Dots, which operate within conversational interfaces. These agents interpret user requests, search for items, and initiate the purchasing sequence. Security measures remain central to the release, requiring active human verification before any money leaves the user's account. This balances the convenience of automation with strict financial controls.

How Checkout Works in ChatGPT Wallet

Users add an eligible card once through Wallet settings or by asking their Dot to open a secure setup form. Their bank may ask them to verify the card during setup. The card saved in Wallet is separate from the one used to pay for a ChatGPT subscription.

At checkout, users see the merchant, what they're buying, the total, and which card will be charged. Each biometric approval creates a record of what the user authorized. The merchant still has to confirm the order.

The setup process requires users to establish their payment preferences ahead of time. Once a card is linked and verified, the agent uses those stored credentials during future interactions. By keeping the subscription billing method separate from the transaction wallet, OpenAI minimizes accounting confusion for subscribers managing multiple digital services.

How Payment Security and Tokenization Protect Users

After approval, the payment network generates a token with its own card number, security code, and expiry date. The payment goes through on the selected card, while the agent and merchant receive the token instead of the user's actual card details. OpenAI says users keep their existing card rewards and protections for eligible disputes, such as those involving defective goods.

Tokenization is a standard cybersecurity practice designed to shield sensitive financial data from potential exposure. By replacing primary account numbers with unique digital tokens, the system ensures that merchants never store actual credit card details. This reduces the risk of financial data theft during online transactions.

Cardholder protections remain intact because the underlying credit card account still backs the purchase. Consumers retain standard rights regarding chargebacks, fraud liability, and rewards accumulation. Financial institutions treat these tokenized transactions similarly to standard online purchases made directly by a human.

Who Can Use ChatGPT Wallet Right Now

OpenAI built Wallet with American Express, Mastercard, and Visa. The initial rollout is limited to US ChatGPT Pro subscribers aged 18 or older who have a US billing address and an eligible US-issued American Express, Chase, Mastercard, or Visa card. Chase customers must be the primary cardholder.

Wallet works in Dots conversations. Regular ChatGPT chats, Work, and standalone Codex threads are outside the initial rollout.

Limiting the initial release to Pro subscribers in the United States allows the organization to monitor system performance and security under controlled conditions. Requirements such as age verification and primary cardholder status ensure compliance with banking regulations and risk management frameworks established by major financial networks.

Merchant Integration and the Commerce Partner Playbook

Some websites restrict or block AI agents, which can prevent checkout. OpenAI published a Commerce Partner Playbook to help merchants and payment providers recognize legitimate purchases made by agents and process them through their existing checkout systems.

Many online retailers use automated bot detection software to prevent unauthorized scraping or fraudulent checkout attempts. Because AI shopping agents interact with websites in unconventional ways, they can inadvertently trigger these security blocks. The partner playbook assists merchants in updating their systems to distinguish between malicious bots and authorized consumer assistants.

Payment processors also play a crucial role in adapting to automated commerce. Ensuring that tokenized transactions flow smoothly through established payment gateways requires cooperation between artificial intelligence developers, financial institutions, and online merchants.

Future Updates and Advanced Controls for Automated Buying

OpenAI expects users will eventually be able to approve purchases in advance by setting spending caps, eligible purchase categories, or time limits. For now, every purchase requires separate approval.

Implementing pre-approval rules will require sophisticated permission structures within the application. Users will need granular control over how much money an automated agent can spend without direct supervision. These upcoming features aim to make routine purchases, such as grocery restocks or standard travel bookings, entirely hands-free.

As the technology matures, observation of user behavior will guide further enhancements. Financial safety, regulatory compliance, and user trust will dictate how quickly these broader automation features become available to a wider audience.