OpenAI releases guide for retailers to enable AI shopping agents
The new handbook helps e-commerce businesses integrate with ChatGPT's agentic features to handle autonomous purchases and inventory checks.
The short version
- OpenAI released a commerce handbook on November 13, 2026, to help retailers support autonomous AI agents.
- Retailers are encouraged to verify their business identity and share real-time pricing and inventory data.
- The guide highlights the need to adjust firewalls and bot filters to prevent blocking legitimate AI transactions.
- Signifyd predicts that agent-led purchases will have 30% higher average values than traditional methods.
- Major platforms like Shopify and Wayfair are adopting agents, while Amazon and United Airlines currently block them.
Why it matters
This marks a shift toward a web where AI agents, rather than humans, perform the heavy lifting of searching and buying. For businesses, it means a new technical standard for 'machine-readable' commerce is becoming mandatory to stay competitive.
A roadmap for agentic commerce
OpenAI recently published a comprehensive guide designed to help online retailers prepare for a new era of autonomous shopping. This handbook provides the technical and organizational steps needed for platforms to interact with ChatGPT's agentic features. These features allow AI agents to find products, check stock levels, and complete purchases on behalf of users.
The initiative comes as OpenAI reports that ChatGPT now reaches 1.2 billion weekly users. By providing a clear path for integration, the company is attempting to turn those users into active shoppers who use AI as their primary personal assistants. The goal is to move beyond simple search results toward a system where the AI can execute the entire transaction flow.
Retailers who adopt these standards can potentially see significant gains in transaction value. According to projections from Signifyd, the average value of purchases made by AI agents is expected to be 30% higher than those made through traditional shopping methods. This suggests that AI agents may be better at identifying and selecting higher-value items for their users.
Technical requirements for shops
To participate in this ecosystem, retailers must first claim and verify their business identity within the ChatGPT environment. The guide specifies that companies need to provide structured product data, including up-to-the-minute pricing and current inventory status. This ensures that the AI agents are making decisions based on accurate, real-time information rather than outdated web scrapes.
OpenAI suggests using plugins to manage shopping carts and verify availability directly. While some browser-based visits by agents do not require specific plugins or data uploads, the guide emphasizes the importance of product feeds and specific interfaces like WebMCP. These tools allow for a deeper level of interaction than a simple web crawl can provide.
The guide also clarifies that specific integrations for digital wallets are not strictly required at this stage. Instead, the focus remains on the flow of product information and the security of the connection. By standardizing how data moves between the store and the agent, OpenAI aims to create a smoother experience for the end user who wants a seamless checkout.
Overcoming security hurdles
One of the most significant challenges for retailers is managing security systems that might mistake an AI agent for a malicious bot. OpenAI identifies firewalls, bot filters, and fraud detection systems as the weakest links in the current process. These systems often block legitimate automated purchases because the transaction patterns look unusual compared to human behavior.
The handbook advises merchants to configure their Content Delivery Networks (CDNs) and protection layers to recognize signed requests from AI agents. To facilitate this, OpenAI recommends using the Web Bot Auth standard. This allows the retailer to distinguish between a harmful scraper and a helpful shopping agent authorized by a human user.
To further reduce the rate of canceled or rejected orders, OpenAI suggests collaborating with established payment and risk management providers. The guide specifically names companies like Adyen, Checkout.com, and Stripe, as well as risk platforms such as Forter, Riskified, Sardine, and Signifyd. These partners help ensure that the novel transaction patterns of AI agents do not trigger false alarms.
Regulations and responsibilities
While the technology allows for more efficient shopping, it also introduces complex regulatory requirements. The guide notes that businesses must navigate the legal framework of the EU AI Act, which governs how these systems can operate. Retailers are encouraged to use a provided implementation guide to understand which AI applications are classified as high-risk and what specific deadlines they must meet.
Despite the automation, the operational responsibility for the sale remains entirely with the merchant. OpenAI clarifies that the seller is still fully responsible for pricing, shipping, fulfillment, returns, and customer service. The AI agent acts as the intermediary for the purchase, but the human-led business must still manage the physical and logistical side of the commerce.
The retail landscape remains divided on how to approach these autonomous systems. While companies like Shopify and Wayfair are moving forward with AI agent integration, others are taking a more restrictive stance. For example, Amazon continues to block purchases made via automated agents, and United Airlines has also denied access to such systems.
What to watch
- How many major retailers adopt the Web Bot Auth standard by the end of the year.
- Whether more major travel and booking platforms follow Amazon's lead in blocking AI agents.
Try it
Check your website's current bot-blocking settings to see if they might accidentally block legitimate automated requests from AI services.
Sources and links
- ReportingBornCity: the article we sourced
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.
Listen to the AI Wire headlines
Listen to the AI WireListen to our daily AI shows every morning: on your commute, in the shower, wherever you are
Listen to the daily showsFirst reported by BornCity · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
