What's Happening?
DoorDash CEO Tony Xu has indicated that third-party AI agents have generated only several hundred orders since their operation began. This figure is notably small when compared to DoorDash's total of 993 million orders reported in the third quarter. The
data suggests that personal AI agents currently account for significantly less than 1% of the platform's overall traffic. This observation aligns with similar trends seen in other major platforms, such as Uber, where AI agent-originated transactions were also a tiny fraction of total activity. For instance, Uber's rides business fulfilled only four agent-originated trips during Halloween week, out of approximately 320 million combined Uber and Uber Eats trips and orders. The distinction is drawn between native AI integrations, which have shown more significant adoption (like OpenTable's OpenAI integration), and third-party agents, which have yet to make a substantial impact on order volumes for these large-scale services.
Why It's Important?
The minimal impact of third-party AI agents on DoorDash's order volume, as reported by CEO Tony Xu, highlights a critical distinction in the adoption of AI technologies within consumer services. While native AI integrations, such as OpenTable's ChatGPT reservation flow, are demonstrating rapid growth and significant user engagement, external AI agents are struggling to gain traction. This divergence suggests that consumers are more comfortable with AI tools directly integrated into a service's platform, where the interaction is often a human-confirmed chat commerce, rather than delegating tasks to independent AI assistants. For businesses, this implies that investment in seamless, in-platform AI experiences may yield greater returns than relying on external AI agent ecosystems. The current low volume of third-party agent activity also indicates that the broader market adoption of personal AI assistants for transactional tasks is still in its nascent stages, posing questions about the future strategies for customer experience leaders in integrating or responding to these emerging technologies.
What's Next?
Despite the current low volume of third-party AI agent activity, businesses are advised to prepare for a potential future increase in their use. Customer experience leaders need to establish clear policies regarding how their systems will interact with personal AI agents. This includes defining acceptable authentication methods, determining the scope of authority for AI agents, and deciding which decisions their systems can safely make when interacting with these agents. While the immediate impact is negligible, the comfort consumers are developing with chat commerce could eventually translate into greater acceptance of personal AI assistants for more complex tasks, such as resolving disputes or canceling services. Companies should monitor the volume of AI agent interactions and revisit customer journeys, especially those with high friction, to ensure they are prepared for a potential rapid scale-up of AI agent activity. The current period of low adoption offers an opportunity for proactive policy development and system adjustments.
Beyond the Headlines
The disparity between the success of native AI integrations and the limited impact of third-party AI agents raises deeper questions about trust, control, and the evolving nature of consumer-AI interaction. Consumers appear to prefer direct, human-like chat interfaces for tasks like reservations, where they maintain a degree of oversight and confirmation. In contrast, delegating tasks entirely to a third-party AI agent, which acts on their behalf, introduces a layer of abstraction that many may not yet be comfortable with, especially for financial transactions or sensitive personal data. This suggests that the ethical and legal frameworks surrounding AI agent autonomy and accountability will become increasingly important as these technologies mature. Furthermore, the development of robust authentication protocols and clear guidelines for AI agent interactions will be crucial to prevent potential misuse or security breaches. The current landscape underscores a cautious, incremental adoption of AI in consumer-facing roles, with a clear preference for integrated, transparent, and user-controlled AI experiences.













