Major retailers and travel brands ban AI agents from browsing and buying

The rise of AI shopping agents

The landscape of online commerce is shifting as AI agents become more capable of performing complex tasks. These bots can now browse websites, build shopping carts, and even complete purchases on behalf of users. This trend has gained momentum following the release of tools like Meta's Muse and personal assistants like Instinct.

Popular chatbots such as Claude and ChatGPT have also integrated agentic capabilities into their systems. These features allow the models to act as personal assistants that can navigate the internet to find products or book services. This shift toward automation has prompted a reaction from the brands that host these digital storefronts.

While the technology offers convenience for consumers, it creates a new set of challenges for businesses. Companies are beginning to reckon with how these bots interact with their platforms. Many are now implementing specific rules to control how, or if, an AI can interact with their customers and data.

Retailers set firm boundaries

Major retail giants are taking a proactive stance by banning unauthorized AI agents from their sites. Amazon, for example, recently showed a popup to users attempting to use Muse for shopping. The message stated that continued access by an unauthorized agent violates the company's Conditions of Use.

eBay has also updated its user agreement to shut the door on AI shoppers. The marketplace specifically prohibits 'buy-for-me agents,' 'LLM-driven bots,' and any end-to-end flows that attempt to place orders without a human reviewing them. These rules are designed to ensure that a person is always in the loop during a transaction.

Target is taking a similar approach by keeping a tighter guest list for AI shoppers. The company's terms prohibit agents from gathering data unless they are explicitly authorized. Only specific partners, including Target's own AI Shopping Assistant, OAI, Google, and Microsoft, are currently permitted to sign in to accounts and place orders.

Travel and booking restrictions

The travel industry is also implementing restrictions to manage security and regulatory risks. United Airlines announced that it is restricting access for third-party agents to prevent unintended purchases. The airline noted that certain actions, like checking in for a flyer, could violate federal regulations requiring a human to acknowledge that bags do not contain banned items.

Booking.com has established a policy where automated assistants cannot make reservations without express permission. The company's terms state that AI-powered assistants are not allowed to access the platform for any reason without prior authorization. They have stated they will block anyone they suspect of using these tools without permission.

Resy, the restaurant booking platform, is taking a more selective approach to AI integration. While the company supports discovery and booking through ChatGPT and Claude, it can block AI activity from other services. They cited concerns about excessive demand on their systems and warned that users violating these rules could have their accounts terminated.

Privacy and traffic concerns

The primary drivers behind these bans are rooted in security and privacy. Companies are worried about how AI agents might access sensitive user data or interact with internal systems in ways that could be exploited. By restricting these bots, brands can better protect their customers from potential data breaches or unauthorized access.

There is also a significant concern regarding the impact on website traffic. If AI agents perform all the browsing and buying, companies fear that fewer humans will visit their sites directly. This could lead to a decline in the organic engagement that many online businesses rely on to stay viable.

Furthermore, the risk of unintended purchases remains a major hurdle for automation. If an AI agent makes a mistake or interprets a command incorrectly, it could lead to costly errors for the consumer. By requiring human oversight, companies aim to ensure that every transaction is intentional and authorized by the account holder.

What comes next for users

For the average consumer, this means that the 'set it and forget it' model of AI shopping may face frequent interruptions. Users may find that their favorite AI tools are blocked from certain websites, requiring them to manually complete the final steps of a purchase or booking. This creates a hybrid experience where AI assists but does not fully execute.

Companies like Yelp are already carving out specific paths for these tools through paid data-licensing programs. These programs allow agents to access data and take actions through approved integrations, such as a specific tool for ChatGPT users to make reservations. This suggests that the future may involve a 'walled garden' approach where only certain bots are allowed to interact with certain brands.

Moving forward, the tension between automation and platform security will likely define the user experience. As AI agents become more sophisticated, brands will need to balance the convenience of these tools with the necessity of protecting their data and maintaining human-centric traffic. Consumers should expect to see more explicit permissions required before an AI can act on their behalf.