OpenAI Expands Shopping Offering with Instant Checkout in ChatGPT, Agentic Commerce Protocol

OpenAI has officially announced the launch of Instant Checkout within ChatGPT, along with the open-source Agentic Commerce Protocol (ACP) that powers it. This new feature will enable users to purchase products directly from supported merchants within the ChatGPT experience, eliminating the need to visit the merchant’s website. Stripe, Etsy, and Shopify were among the initial partners, with Etsy purchases supported from day one for U.S. users. Shortly after the launch, Walmart announced a partnership with OpenAI, including an integration with Instant Checkout. This marks the latest and perhaps most significant step OpenAI has taken toward becoming a shopping destination.

Instant Checkout enhances the current shopping experience within ChatGPT, with Instant Checkout items displayed alongside traditional items that link to a merchant’s website. The feature will be free for users, while merchants will incur a fee for each completed purchase. 

OpenAI states that results are not sponsored and that neither an item’s Instant Checkout status nor the merchant fee will affect product rankings. However, the Instant Checkout status will be a “considered factor” in ranking multiple merchants selling the same item, according to an OpenAI announcement. Merchants can choose to apply only for inclusion in ChatGPT’s shopping search results or for both search results and the Instant Checkout program if they are committed to fully integrating the ACP with their checkout and payment systems.

The Agentic Commerce Protocol was co-developed by OpenAI and Stripe, and as an open-source protocol, OpenAI aims for it to become the “open standard for AI commerce.” With the immediate release of the ACP, merchants can now begin integrating it into their checkout systems. 

This protocol enables merchants to maintain complete control over payments, systems, and customer relationships; however, integrating the technology into their checkout and payment systems requires a significant effort. OpenAI requires endpoints to start and update a checkout session with items, fulfillment options, and buyer details, as well as an API endpoint with payment options to complete the checkout. Notably, Stripe is the only natively supported payment provider at launch.

OpenAI’s Instant Checkout offers another way to reduce buying friction for consumers, making it easier than ever to go from consideration to purchase. For brands, this solution is a double-edged sword. As the newest off-site shopping experience, Instant Checkout provides an additional platform to showcase and sell its products. 

Still, it also raises concerns about losing control of the user’s shopping journey, consumers spending less time on brands’ owned platforms, and the potential loss of marketing signals during direct on-site visits and purchases. While there will clearly be trade-offs for brands that embrace OpenAI’s vision for AI commerce, the impact of AI on consumers’ shopping habits cannot be ignored. According to a June report from Statista, nearly 60% of shoppers used ChatGPT for inspiration, making this new offering difficult to overlook, and the potential benefits highly attractive.

Understanding chat behavior with AI has become a top priority for marketers, but gaining insight into AI responses is essential for identifying a brand’s agentic commerce opportunities. Recent user data published by OpenAI in collaboration with the National Bureau of Economic Research showed that most ChatGPT conversations do not explicitly indicate shopping intent. However, AI responses to discussions about practical life problems often pave the way for agentic shopping experiences. 

At a recent industry event on October 8, new research shared by Profound, an AI Visibility Measurement provider, revealed that 33% of AI conversations included unprompted product recommendations. Additionally, nearly half of these conversations featured AI mentioning brand names without being prompted. As search shifts toward AI-driven experiences, brands must focus on understanding their visibility in conversational interfaces, where AI will recommend products and customers will make purchases. Success will depend on understanding visibility in AI conversations, maintaining data quality, and seamless integration with agentic protocols like ACP.

Brands will also need to consider the technical effort required to integrate the Agentic Commerce Protocol into their systems in order to enable Instant Checkout. Although brands can opt out of Instant Checkout, their products will still appear alongside checkout-enabled items in ChatGPT’s product-based search results, using a fairly standard product feed. Notably, we’ve seen similar API-based off-site checkout concepts fail to succeed in the marketplace, such as Meta’s recently discontinued checkout on Facebook and Instagram initiative, which has been adjusted to only support adding items to the cart on Facebook and Instagram, with users ultimately directed to the merchant’s site to complete their purchase. 

Brands will also need to ensure that the tool’s payment protocol is properly integrated with their current checkout systems to maintain PCI compliance. We recommend that brands maintain transparency about any partners involved in the process and who have access to customer data, ensuring necessary compliance with data regulations.

OpenAI’s ACP is currently focused on shopping, but similar integrations are likely to expand beyond the retail sector. Brands across travel, hospitality, events, and other industries with significant feed and Google structured data integrations should prepare for a future where agentic protocols serve as gateways to booking experiences.

Looking ahead, brands will need to adopt a comprehensive total search strategy for structured data sooner than later to ensure that essential details, such as product prices (including sales and promotions), are consistent across ChatGPT listings, Google Merchant listings, and product pages.

Agent-focused content will become increasingly important in shaping AI product recommendations that lead to agentic shopping opportunities. Aligning product content with query fan-out—a process in which a user’s single query (or prompt) is split into multiple sub-queries to collect information from various sources—is an agent-friendly approach to providing training data for AI recommendations. 

Similar to the emergence of commerce platforms like Amazon, brands may need to give up some control over how their products are showcased in AI experiences. This might be less of a concern for categories like consumer packaged goods, but premium brands in luxury, fashion, and beauty are likely to have significantly less influence over key stylistic elements that drive conversion on the brand website shopping experience.

For brands weighing their options, there are still many unknowns across these developments.

OpenAI states that currently, all results are organic and that checkout-enabled items are not favored in search results. However, in December last year, OpenAI told the Financial Times that it was considering its options for ads across its services and technologies. The Financial Times also noted that OpenAI has been hiring advertising talent. Coupled with the latest technological developments, these signals suggest a future in which OpenAI could add sponsored listings to search results, adjust the recommendation algorithm to favor checkout-enabled items, or introduce new experiences limited to paid placements.

There remain questions around the open-source nature of the ACP, which OpenAI positions as the future of AI commerce—not just within ChatGPT but across all platforms and partners in the “agentic commerce” landscape. If other companies and AI solutions adopt ACP and OpenAI succeeds in establishing it as a standard, the impact of this launch could be much more significant.

With the potential for future “pay to play” elements in ChatGPT results and the possibility that the ACP could become an industry standard, it may be in a brand’s best interest to adopt this technology early to avoid being left behind later.