# DoorDash opens waitlist for a text-to-order AI agent in Apple Messages The delivery company says its new agent can interpret prompts such as "order my usual," suggest carts from local restaurants and handle group orders with mixed dietary preferences, with a U.S. waitlist now open. Canonical URL: https://freelancenews.online/news/doordash-opens-waitlist-for-a-text-to-order-ai-agent-in-apple-messages-92c62720 Published: 2026-10-01T02:17:35.648Z Updated: 2026-10-01T02:17:35.648Z Source published: 2026-09-30T16:00:24.000Z Event date: Not established Review status: source-reviewed Review method: Automated comparison against retrieved source text; not independent fact-checking. ## Report DoorDash announced on Wednesday that it is launching a text-to-order AI agent that lets users place food orders through Apple Messages, according to TechCrunch. The company is opening a waitlist for users in the United States to try the feature, so access is not immediate or general at announcement time. The mechanism described is conversational rather than app-based. A user sends a prompt such as "order my usual," and the agent is said to understand that this refers to their Friday night order. That example implies the agent leans on a user's prior ordering pattern rather than requiring a full item-by-item specification each time. Beyond reordering, DoorDash says users can ask for a specific dish or request a local recommendation. The agent then searches local spots, suggests a cart based on the prompt, and can text photos of the food it recommends. Those are the company's claims about behavior; the supplied material does not include independent testing of accuracy or how recommendations are ranked. Another capability that has been documented is group ordering. Mixed dietary preferences, along with varying quantities, can be handled by the agent within one order, according to DoorDash. Compared with a single-user reorder, that case is meaningfully harder, since conflicting constraints across multiple people must be reconciled before a cart is finalized. An edge over rivals Uber Eats and Grubhub is what DoorDash is looking to gain, and that competitive framing is stated explicitly. A broader push toward personal AI agents, which handle tasks without users navigating apps or websites themselves, is also the context in which the launch is placed. That context is the source's characterization, not a measured market shift. Separately from the agent, DoorDash announced it will begin testing delivery drones with select restaurants in Northern California. The evidence does not specify which restaurants, how many, what the test geography covers beyond the region, or when the testing starts, so those details remain unknown. For freelancers, designers and developers, the practical read is about where ordering interfaces are heading. If a text thread can assemble a cart, the surface a restaurant or brand controls shifts from an app screen to a conversational exchange, and the agent's interpretation of an ambiguous request becomes the effective product experience. That shift carries tradeoffs the announcement does not resolve. A conversational agent that infers intent from history may be convenient for repeat orders but opaque when it guesses wrong, and the supplied material says nothing about how users correct a bad cart, how recommendations are surfaced as sponsored or organic, or what data the agent draws on. There is also no pricing information in the evidence. It is not stated whether the agent carries fees, changes delivery pricing, or alters how restaurants appear in results. Without that, any cost comparison against ordering through the existing app would be speculation. Availability is the clearest limitation. The feature is gated behind a U.S. waitlist, and the evidence gives no rollout timeline, no indication of which markets are first, and no statement about platform requirements beyond Apple Messages. Non-U.S. users and Android users have no documented path to access at this point. The drone testing adds a second, distinct workstream. Because it is described only as testing with select restaurants in Northern California, it should be treated as an early experiment rather than a service change. The evidence does not connect the drone test to the AI agent in any operational way. What remains unknown is substantial: no accuracy figures, no described evaluation method, no sample of users, and no independent verification of the ordering, recommendation or group-order claims. Everything about the agent's behavior in this report comes from DoorDash's own description as relayed by TechCrunch. A reasonable conclusion for this audience is to watch the waitlist rollout as a signal about conversational commerce rather than to plan around it. The concrete, supported facts are narrow: an announced agent, a U.S. waitlist, described capabilities, a competitive motive, and a separate Northern California drone test. ## Key points - DoorDash announced a text-to-order AI agent that works through Apple Messages, with a U.S. waitlist now open. - The company says the agent can interpret prompts like "order my usual," suggest carts from local spots, send food photos, and handle group orders with mixed dietary preferences. - DoorDash frames the launch as an effort to gain an edge over Uber Eats and Grubhub, within a broader push toward personal AI agents. - Separately, DoorDash said it will begin testing delivery drones with select restaurants in Northern California. - No pricing, rollout timeline, accuracy data or independent testing is included in the supplied evidence. ## Practical implications — editorial interpretation Editorial interpretation: if ordering migrates into a text thread, the agent's interpretation of an ambiguous request becomes the de facto interface that restaurants and brands are designing for, which matters to anyone building ordering, menu or recommendation experiences. The evidence does not show how merchants control their presentation inside that agent, so treat this as a direction to monitor rather than a channel to build against today. ## Limitations and unknowns Access is limited to a U.S. waitlist with no stated timeline or market order. The evidence contains no pricing, no accuracy or evaluation data, no user sample, and no independent verification; all capability claims come from DoorDash as relayed by TechCrunch. The drone test's restaurants, scale and start date are unspecified, and no link between the drone test and the AI agent is documented. ## Sources - [1] techcrunch.com: DoorDash launches an AI agent you can text to order food https://techcrunch.com/2026/09/30/doordash-launches-an-ai-agent-you-can-text-to-order-food/ Retrieved: 2026-10-01T02:17:24.405Z ## Claim references - DoorDash announced a text-to-order AI agent that lets users order through Apple Messages. [source 1] - The agent is described as understanding prompts such as "order my usual" as a reference to a user's Friday night order. [source 1] - Users can ask for a specific dish or a local recommendation, and the agent searches local spots and suggests a cart, including texting photos of recommended food. [source 1] - DoorDash says the agent can handle group orders with mixed dietary preferences and different quantities in one order. [source 1] - DoorDash is opening a U.S. waitlist for the feature. [source 1] - DoorDash also said it will begin testing delivery drones with select restaurants in Northern California. [source 1]