DoorDash’s AI-agent revolution has produced only a few hundred orders
DoorDash’s AI-agent revolution has produced only a few hundred orders
DoorDash began laying groundwork for AI-assisted ordering well before the latest numbers landed. It launched its Ask DoorDash chatbot in June, followed with a command-line agent in July and a text-to-order tool in September. Its tools are aimed at making choices and cart-building easier inside DoorDash’s own ecosystem, rather than handing the customer relationship to outside bots.
Then, less than a month before its third-quarter earnings call, DoorDash introduced a Model Context Protocol connector for selected early beta partners, chiefly to streamline corporate meal orders. The connector offers third-party AI tools a standardized route into the platform—but the early traffic has been microscopic.
On the call, CEO Tony Xu said outside agents such as Muse, Instinct and Fo had produced only “maybe several hundred orders in total.” That is a rounding error against DoorDash’s roughly 3 million daily deliveries, and Xu said consumers still overwhelmingly order through the DoorDash app.
Xu’s message was neither dismissal nor surrender to the AI hype cycle. “You never want to fight gravity,” he said. “But at the same time, you don’t want to be delusional.” He argued agents may prove more useful for high-consideration purchases or decisions requiring background research; food delivery, by contrast, remains a quick, direct transaction.
There is a more optimistic counterpoint within DoorDash’s own data. The company says orders using Ask DoorDash have higher basket values and more items than traditional orders, suggesting AI may reshape ordering from inside the app before outside agents gain real traction. DoorDash is also leaving the door open to partnerships—and potentially advertising—inside third-party agents if their audiences finally arrive.
Write a comment