DoorDash CEO reveals outside AI agents drove only hundreds of orders

Outside artificial intelligence agents have generated only a few hundred orders on DoorDash since launch, according to company co-founder and CEO Tony Xu. Speaking on a third-quarter earnings call, Xu characterized this volume as a very low use case for a platform handling nearly 1 billion total orders in the same period.

This tiny tally arrives despite widespread industry enthusiasm for LLMs taking over routine consumer purchases and digital shopping tasks. Our read is that while third-party automated tools struggle to gain traction for simple chores, native platform features tell a different story. DoorDash reports that its internal Ask DoorDash tool builds baskets five times faster than traditional methods, with basket values averaging nearly 50 percent higher.

These internal tools succeed by guiding users through recipe photos or vague prompts rather than acting as fully autonomous external shoppers. Technical hurdles also explain the friction. DoorDash rolled out an open-standard Model Context Protocol connector less than a month ago to let third-party AI tools interface directly with its platform.

That capability remains restricted to a small group of early beta partners mostly handling corporate lunch orders.

Without native integrations, outside agents must navigate standard web interfaces.

Xu cautions against fighting technological gravity while rejecting delusion about current metrics.

Whether autonomous agents eventually dominate food delivery or remain a novelty, the early data suggests consumers still prefer ordering directly for basic tasks.