What's Happening?
DoorDash CEO Tony Xu has indicated that third-party AI agents are generating only a negligible number of orders for the food delivery platform. During an October 29 earnings call, Xu stated that the company
has observed 'maybe several hundred orders in total' from these external AI agents. This cumulative figure is remarkably small when compared to the 993 million orders DoorDash processed in the third quarter alone, which marked a 28% increase. This data suggests that despite integrations with AI platforms like ChatGPT for food ordering since 2025, customer adoption of these third-party AI agents for DoorDash services remains extremely low. The company's shareholder letter also noted the deployment of a Model Context Protocol server to interface with external AI agents, including those used for corporate ordering, yet these integrations have not translated into significant order volume. This trend mirrors similar observations from other companies in the gig economy, indicating a broader challenge in driving consumer engagement with AI-driven ordering systems for habitual purchases.
Why It's Important?
The minimal order volume generated by third-party AI agents for DoorDash highlights a significant challenge for the broader e-commerce and AI industries. Companies have invested in integrating with AI assistants, anticipating a new channel for customer engagement and sales. However, DoorDash's experience, alongside similar reports from Uber, suggests that consumers are not readily adopting these AI agents for routine transactions like food delivery. This could lead to a re-evaluation of AI integration strategies, with businesses potentially shifting focus from third-party agents to developing in-app, first-party AI assistants that offer a more seamless and integrated user experience. The lack of traction also indicates that the perceived convenience of AI agents may not outweigh the established habits of direct app usage for frequent, low-consideration purchases. This could impact investment decisions in AI development and partnerships, as companies may become more cautious about allocating resources to external AI platforms that do not demonstrate a clear return on investment.
What's Next?
In response to the low adoption rates of third-party AI agents, DoorDash and similar companies may pivot their AI strategies. Instead of relying on external platforms, there could be an increased focus on developing and enhancing proprietary AI tools within their own applications. DoorDash's shareholder letter mentions 'Ask DoorDash,' which assists with grocery basket building and restaurant suggestions, and 'Text DoorDash,' a texting channel for ordering. These internal tools aim to improve the user experience directly within the DoorDash ecosystem. The company's strategic focus on 'winning the battle for atoms'—referring to its physical delivery network—suggests that while it acknowledges the potential of AI, its primary investment will remain in logistics and operational efficiency. Future developments may involve more sophisticated in-app AI features that leverage user data for personalized recommendations and streamlined ordering, rather than pushing users to interact with third-party AI agents. This shift could also influence how AI agent platforms evolve, potentially prompting them to offer more compelling value propositions to attract both businesses and consumers.
Beyond the Headlines
The limited success of third-party AI agents in driving orders for DoorDash points to a deeper behavioral insight: consumers often prioritize established habits and direct control over perceived technological advancements for routine tasks. While AI agents promise efficiency, the existing user interfaces of delivery apps are already highly optimized for quick, habitual purchases. This suggests that for high-frequency, low-consideration transactions, the friction introduced by an intermediary AI agent might outweigh its benefits. This trend could also have implications for data privacy and user trust; consumers might be more comfortable sharing their preferences directly with a known service provider rather than through a third-party AI. Furthermore, the struggle of AI agents to gain traction in this sector could indicate that the true value of AI in e-commerce might lie in more complex tasks requiring extensive comparison, research, or assembly, such as grocery list compilation or travel planning, rather than simple reordering of familiar items. This could lead to a bifurcation in AI application, with different types of AI tools succeeding in different market segments based on the complexity and frequency of the user's needs.
