Tuesday, November 10, 2026

Apps & Consumer

DoorDash: AI agents have placed only 'several hundred' orders

DoorDash's CEO said AI shopping agents have so far generated maybe several hundred orders in total, compared with the company's average of 3 million deliveries a day.

Hand holding smartphone with food delivery app on screen in bright setting.
Photo: Pexels / Erik Mclean

DoorDash CEO Tony Xu said during the company’s third-quarter earnings call that AI agents built by third parties — including Muse, Instinct, and Fo — have placed only maybe several hundred orders in total on the platform so far. That’s a fraction of DoorDash’s average of 3 million deliveries a day, per its 2025 report. Asked about agentic ordering, Xu said: “I would say that this is something that is still a very dynamic area … you never want to fight gravity. But at the same time, you don’t want to be delusional.”

Xu said users still place most orders through DoorDash’s own app, but that the company wants to understand agentic use cases better and build a working relationship with the makers of these third-party agents. He suggested DoorDash could eventually advertise inside the agents themselves. The company has also built its own agents: a text-to-order agent launched in September, a command-line agent introduced in July, and “Ask DoorDash,” a chatbot launched in June that takes orders through prompts and photos. It has also released an MCP — a protocol letting other chatbots place orders — aimed at corporate ordering.

The rules of agent-based ordering are still being worked out: some retailers are blocking AI shopping agents from accessing their sites, while others are opening up. Amazon has blocked these personal agents, while Walmart has welcomed them, looking to pull business away from rivals.

Why it matters

DoorDash’s numbers suggest agentic commerce remains a novelty rather than a real sales channel today, even as the company builds out its own suite of agents and keeps the door open to advertising inside third-party ones. How platforms treat these agents — blocking them like Amazon or welcoming them like Walmart — could still reshape who captures a fast-growing new type of transaction.