DoorDash said it’s only received ‘several hundred’ orders from AI agents

AI Agents in Food Delivery: The Current Reality

DoorDash’s latest earnings call revealed that, despite the hype surrounding AI‑driven commerce, the company has processed only “several hundred” orders from external AI agents such as Muse, Instinct, and Fo. By contrast, DoorDash handles roughly 3 million deliveries each day—a scale that makes the modest AI‑order count a drop in the bucket.

Why the Numbers Are So Low

• User habits remain app‑centric – Most customers still reach for the DoorDash mobile app or website to place orders. The convenience of a familiar interface outweighs the novelty of an AI intermediary.
• Technical and policy barriers – Many e‑commerce sites are still deciding whether to expose their ordering APIs to third‑party agents. Amazon, for example, blocks such access, while Walmart is more permissive, creating an uneven landscape for AI agents.
• Early‑stage technology – AI agents are still experimenting with natural‑language ordering, image‑based prompts, and integration with corporate procurement tools. The technology is evolving, but it has not yet reached a point where it can reliably replace a human‑initiated order.

DoorDash’s Own Agent Initiatives

While external agents remain limited, DoorDash is actively building its own AI ordering tools:

1. Text‑to‑order agent (launched Sep 2025) – Users can send a simple SMS to place an order without opening the app.
2. CLI agent (July 2025) – A command‑line interface for power users and developers to script orders.
3. Ask DoorDash chatbot (June 2025) – A conversational bot that accepts prompts and photos, turning visual cues into menu selections.
4. MCP for corporate ordering – An integration that lets businesses route orders through internal chat platforms.

These internal agents give DoorDash direct control over the user experience and data, positioning the company to experiment with monetization options such as advertising within the AI flow.

Looking Ahead: Opportunities and Risks

• Potential ad revenue – If AI agents become a common ordering channel, DoorDash could sell ad placements inside the agent’s conversation, similar to sponsored results in search engines.
• Partnerships with third‑party agents – By offering APIs or revenue‑share models, DoorDash could encourage more agents to integrate its catalog, expanding reach beyond the app.
• Regulatory scrutiny – As AI agents handle transactions, consumer‑protection and data‑privacy regulations may tighten, requiring transparent disclosures about who is placing the order.
• Competitive dynamics – Retail giants that already block agents (e.g., Amazon) could limit growth, while more open platforms (e.g., Walmart) may become testing grounds for AI‑driven commerce.

Practical Takeaway

For now, AI agents represent a niche experiment rather than a mainstream ordering method. DoorDash’s strategy of building its own agents while monitoring third‑party activity allows it to stay agile without overcommitting resources. Marketers and product teams should watch the space for incremental adoption, but continue to prioritize the core app experience where the vast majority of orders still occur.