What's Happening?
DoorDash CEO Tony Xu has reported that the company has received only 'maybe several hundred orders in total' from external AI agents since their introduction. This statement aligns with a broader trend observed across major booking platforms, where AI agent-initiated
bookings remain a very small fraction of overall transactions. Uber reported only four agent-booked rides in its busiest week, while Airbnb's CEO Brian Chesky described AI agent bookings as 'very non-material.' Booking Holdings indicated that AI referrals accounted for 'significantly less than 1%' of Booking.com's room nights. Expedia's Ariane Gorin also noted that AI agents have not significantly altered booking methods. The primary reason cited for this low adoption is that the final booking and payment steps for services like hotel rooms and rides often occur outside the AI chat interface, requiring users to navigate to external websites or apps. An exception is OpenTable, which saw 190,000 diners booked through ChatGPT in a month, largely because the booking process is completed entirely within the chat.
Why It's Important?
The limited impact of AI agents on major platforms like DoorDash, Uber, Airbnb, and Booking.com highlights a significant hurdle in the widespread adoption of AI for transactional services. While AI agents are gaining users, their inability to complete transactions seamlessly within the chat environment creates friction, deterring users from finalizing purchases. This indicates that for AI agents to become truly disruptive, they must integrate end-to-end booking and payment functionalities. The success of OpenTable within ChatGPT demonstrates that when the entire process is streamlined and contained, user adoption increases significantly. This trend suggests that companies investing in AI agent technology need to prioritize direct integrations and in-chat completion to capture market share, otherwise, these agents will remain primarily tools for information gathering rather than transaction execution. The current scenario also underscores consumer reluctance to allow AI agents to handle payments, with surveys indicating a strong preference for human interaction in case of issues.
What's Next?
The future of AI agent bookings hinges on the development and implementation of in-chat transaction capabilities. Companies like Expedia and Cloudbeds are working on direct hotel connections to AI agents like Muse, with some expected to go live in the first quarter of 2027. Airbnb plans to roll out its own AI travel agent in the first half of 2027. These developments aim to enable users to complete bookings directly within the AI interface, potentially increasing adoption rates. The first major test of these new integrations will come in February when Booking Holdings and Expedia report their fourth-quarter earnings, which may include data on bookings completed inside AI. Until then, most AI agent interactions will likely continue to serve as initial research tools rather than final booking platforms. Companies are also building their own proprietary AI agents within their apps, aiming to retain direct customer relationships and control the user experience.
Beyond the Headlines
The slow uptake of AI agent bookings reveals deeper implications regarding trust, user experience, and the evolving role of AI in commerce. Consumers' hesitation to allow AI to handle payments, coupled with their preference for human support in case of problems, highlights a fundamental trust deficit that AI systems must overcome. This suggests that while AI can automate many aspects of the customer journey, the human element remains crucial for high-value or complex transactions like travel. Furthermore, the current landscape indicates a strategic battle between platforms and independent AI agents. Platforms are developing their own AI solutions to maintain direct customer relationships, potentially limiting the influence of third-party AI agents. This could lead to a fragmented AI commerce ecosystem where different AI agents are optimized for specific platforms, rather than a universal AI assistant capable of seamless cross-platform transactions. The ethical implications of AI handling personal financial data and the responsibility for transaction errors also remain significant considerations.
