OpenAI on Friday introduced a significant new feature allowing artificial intelligence agents to make purchases on behalf of users directly within ChatGPT through existing payment networks, marking a major step forward in the evolution of conversational commerce. The rollout, developed in collaboration with major financial institutions American Express, Mastercard, and Visa, signals a transformative shift in how consumers may interact with digital platforms and online merchants in the near future.
Under the mechanics of the newly unveiled system, the integrated wallet lets users authorize transactions securely using biometric verification methods, such as facial recognition or fingerprints. To ensure privacy and security, payment networks generate tokenized card credentials for each approved purchase. This crucial technical mechanism keeps users’ actual card details completely hidden from both merchants and the AI agents themselves, mitigating traditional cybersecurity risks associated with online shopping and third-party data exposure.
Initially, the service is being rolled out to ChatGPT Pro subscribers located in the United States who possess eligible cards issued within the country. However, this initial release is merely the first phase of a broader roadmap. OpenAI has stated that it plans to eventually support purchases within preset spending limits and other customized restrictions, which
would eventually eliminate the need for users to manually approve each individual transaction, thereby streamlining automated tasks like booking travel, purchasing routine household goods, or managing subscriptions.
The launch comes at a time when major artificial intelligence companies are aggressively seeking to move beyond mere product recommendations into actually executing transactions. This strategic evolution has the potential to fundamentally change how consumers interact with online merchants, bridging the gap between discovery and purchase within a single conversational interface. OpenAI noted that it is actively working with traditional payment processors and fraud prevention firms to ensure that these AI-driven purchases can integrate smoothly through existing checkout systems without requiring merchants to overhaul their current infrastructure.
With this move, OpenAI joins a rapidly growing and competitive field of companies developing specialized payment systems tailored specifically for AI agents. The race to capture the infrastructure layer of autonomous commerce has attracted heavyweights from both traditional finance and the technology sector. Google introduced its Agent Payments Protocol in 2025, while legacy credit card giants Visa and Mastercard have simultaneously launched their own dedicated initiatives to support AI-initiated transactions, recognizing that automated agents will soon represent a substantial volume of global digital commerce.
Simultaneously, cryptocurrency and blockchain companies are also heavily developing payment infrastructure designed explicitly for AI agents, offering alternative rails for autonomous machine-to-machine transactions. Coinbase introduced its x402 protocol in 2025, enabling autonomous stablecoin payments for digital services without human intervention. Similarly, Circle launched its Agent Stack in May, allowing agents to hold and spend USDC within predefined cryptographic limits. Meanwhile, Crossmint, a company that provides stablecoin infrastructure, has taken a different strategic approach by combining modern stablecoin wallets with traditional card payment networks, attempting to bridge the two disparate financial worlds.
The sheer scale of the emerging market explains the intense corporate interest and rapid product development across the technology and financial sectors. Global agentic commerce transaction value is expected to reach a staggering $1.5 trillion by 2030, according to an April report from Juniper Research. This represents an astronomical leap from an estimated $8 billion in 2026, highlighting the explosive growth trajectory projected for AI-driven economic activity over the remainder of the decade as autonomous agents transition from experimental tools into active economic participants capable of managing complex financial tasks on behalf of human users.





