DoorDash has revealed that third-party artificial intelligence agents have generated only several hundred orders on its platform so far, highlighting the gap between the growing excitement around agentic commerce and its actual adoption by consumers. During its Q3 2026 earnings call on October 29, CEO Tony Xu told analysts that orders from external AI assistants remain limited, with most customers continuing to place food and grocery orders directly through the DoorDash app.
The disclosure comes as technology companies increasingly promote AI agents capable of searching for products, comparing options and completing purchases on behalf of users. However, DoorDash’s experience suggests that consumers have not yet widely adopted AI-powered ordering for everyday food delivery. The company is still exploring partnerships with third-party agents while developing its own AI-powered ordering tools.
Key Takeaways
- AI agent orders: DoorDash has received only several hundred orders from third-party AI agents.
- CEO: Tony Xu described agent-based ordering as an early-stage and developing market.
- Daily deliveries: DoorDash reported an average of more than 3 million daily deliveries in its 2025 report.
- Third-party agents: The company referenced platforms including Muse, Instinct and Fo.
- Own AI tools: DoorDash has introduced Ask DoorDash, Text DoorDash and a Model Context Protocol (MCP) integration.
- Future opportunity: DoorDash sees potential for partnerships, incremental orders and advertising within AI agents.
- Main challenge: Ensuring that orders are delivered accurately and reliably remains more complicated than simply placing an order.
DoorDash Says AI Agents Have Generated Only Several Hundred Orders
DoorDash CEO Tony Xu disclosed the figures during the company’s third-quarter 2026 earnings call.
An analyst asked how the company was preparing for a future in which consumers increasingly use AI agents to purchase food, groceries and other products.
Xu responded that third-party AI agents had generated only several hundred orders in total during the period they had been operating with DoorDash.
He described the use case as relatively small and said customers still predominantly use the company’s own application for everyday food orders.
The comments provide a reality check for the growing discussion around agentic commerce, a concept in which AI systems can independently perform shopping and purchasing tasks instead of simply recommending products.
DoorDash’s figures suggest that, despite increasing interest from technology companies, external AI agents have not yet become a meaningful ordering channel for its platform.
Several Hundred AI Orders Compared With Millions of Daily Deliveries
The difference between AI-generated orders and DoorDash’s traditional delivery business is substantial.
According to the company’s 2025 report, DoorDash handled an average of more than 3 million deliveries per day.
The several hundred orders mentioned by Xu represent a very small volume compared with that established delivery operation.
However, the figures cover different reporting periods. DoorDash did not disclose the precise timeframe behind the several hundred AI-agent orders, so the comparison should be treated as an indication of scale rather than a like-for-like adoption rate.
The company’s third-quarter 2026 results provide another indication of its overall scale.
| Business metric | Q3 2026 |
|---|---|
| Total orders | 993 million |
| Revenue | $4.6 billion |
| Marketplace gross order value | $33.9 billion |
| Adjusted EBITDA | $1.06 billion |
| Net income | $466 million |
These figures cover DoorDash’s overall business and should not be confused with the much smaller number of orders generated through third-party AI agents.
Which AI Agents Are Sending Orders to DoorDash?
DoorDash’s CEO discussed third-party AI assistants, including Muse, Instinct and Fo, in the context of agent-based ordering.
These systems are designed to help users perform tasks through natural-language instructions rather than navigating every application manually.
For example, a user could potentially ask an AI agent to:
- Find a restaurant serving a particular type of food.
- Compare available restaurants and menu options.
- Select items based on dietary preferences.
- Place an order through a connected delivery platform.
- Arrange food delivery without manually navigating multiple screens.
The idea is to make AI agents an additional interface between consumers and online services.
However, the experience depends on whether platforms allow external agents to access their services and whether the ordering process can be completed reliably.
DoorDash is exploring how these partnerships could create additional orders rather than simply shifting existing customers away from its own application.
DoorDash Is Building Its Own AI Ordering Tools
Although third-party AI orders remain limited, DoorDash is actively developing its own AI-powered shopping and ordering experiences.
The company has introduced several tools designed to make ordering easier.
1. Ask DoorDash
Ask DoorDash is an AI-powered assistant that helps customers discover restaurants and build shopping baskets.
Users can describe what they want in natural language or use photographs to communicate their requirements.
For example, a customer could provide a recipe image and ask the assistant to help create a grocery basket containing the necessary ingredients.
The feature is designed to reduce the time spent searching through menus and product listings.
2. Text DoorDash
DoorDash introduced Text DoorDash in September 2026.
The feature allows customers to interact with the service through text-based instructions, making ordering possible through a more conversational interface.
Instead of navigating several screens, customers can communicate what they want through messages.
3. Command-line AI agent
DoorDash also introduced a command-line interface (CLI) agent in July 2026.
The tool provides another way for users and developers to interact with DoorDash’s services through commands and automated workflows.
4. Model Context Protocol integration
DoorDash has also released a Model Context Protocol integration.
MCP is an open standard that allows AI applications to connect with external tools and services.
DoorDash’s integration allows third-party AI systems to interact with its ordering capabilities, including use cases such as corporate lunch orders.
This approach could make it easier for external assistants to access DoorDash without requiring users to manually navigate the company’s application.
Why Are Consumers Still Using DoorDash’s App?
One reason AI-agent ordering has not gained significant traction is that ordering food is already relatively straightforward.
Customers can open the DoorDash app, select a restaurant, choose menu items, pay and track delivery.
Using an external AI assistant may not offer enough additional convenience for a task that already requires relatively few steps.
Tony Xu suggested that AI agents could be more useful for purchases involving longer consideration periods or tasks that users may not be actively thinking about.
For example, an AI assistant could eventually help users plan meals, identify grocery requirements, compare products and arrange purchases proactively.
These activities involve more decision-making than simply ordering a familiar meal.
DoorDash therefore sees a distinction between routine food ordering and more complex shopping experiences where an AI agent could provide additional value.
Delivery Reliability Is More Important Than Placing the Order
Xu also highlighted an important challenge in agentic commerce: completing a purchase is only one part of the customer experience.
For a delivery platform, the actual challenge is ensuring that the order arrives correctly, on time and in the expected condition.
This involves several operational steps:
- Confirming that restaurants or retailers have the requested items.
- Ensuring that orders are prepared correctly.
- Assigning available delivery workers.
- Managing delivery routes and estimated arrival times.
- Handling missing or incorrect items.
- Resolving payment, refund and customer-support issues.
An AI agent may be able to select products and submit an order, but it cannot automatically guarantee that the physical delivery process works as expected.
This is why DoorDash is interested in understanding how external agents can create additional demand while maintaining control over the underlying delivery experience.
DoorDash Sees Advertising Opportunities in AI Agents
Another interesting part of Xu’s comments concerned the potential business opportunities created by third-party AI assistants.
DoorDash generates revenue through its marketplace, delivery services and advertising products.
If consumers increasingly use AI agents to discover restaurants and select products, the way businesses reach customers could change.
Xu suggested that advertising opportunities could eventually emerge within third-party AI agents.
For example, restaurants and retailers might pay to have their products or offers presented to consumers during relevant AI-assisted shopping experiences.
However, this would require partnerships between AI companies and commerce platforms, along with clear rules around recommendations, sponsored placements and consumer choice.
DoorDash has not announced a specific advertising partnership or revenue-sharing arrangement for this market.
The comments indicate that the company is exploring how its existing advertising business could evolve if AI agents become a more common shopping interface.
AI Agent Commerce Faces a Bigger Adoption Challenge
DoorDash’s experience highlights a broader challenge for the emerging agentic commerce industry.
Technology companies are developing AI systems that can search websites, compare products, interact with applications and complete transactions.
But consumers must trust these systems before allowing them to spend money on their behalf.
Several challenges remain:
Trust and security: Users need confidence that agents will select the correct products and handle payment information securely.
Website access: Online platforms must decide whether external agents can access their services and complete purchases.
Payment systems: AI agents need reliable ways to authenticate transactions and obtain user approval.
Customer support: Businesses must determine who is responsible when an agent places an incorrect order.
Commercial relationships: Platforms and AI companies need agreements covering commissions, advertising and customer ownership.
The industry is therefore still working through both technical and commercial questions.
DoorDash’s early figures suggest that widespread agent-driven ordering has not yet materialised in its food-delivery business.
The Bigger Picture
DoorDash’s disclosure provides a useful measure of the difference between the expectations surrounding agentic commerce and its current adoption. Although AI agents are becoming more capable of interacting with digital services, only several hundred third-party agent orders have been recorded by DoorDash so far. Most customers continue to use the company’s established application for routine food orders. The figures do not establish that agentic commerce has failed, but they show that technical capabilities have not automatically translated into large-scale consumer behaviour.
For DoorDash, the opportunity may lie in combining its delivery infrastructure with AI-powered discovery, shopping and ordering experiences. Its own tools and MCP integration show that the company is preparing for customers to interact with its services through different interfaces. The longer-term challenge will be making AI-generated purchases genuinely incremental while protecting delivery quality, advertising revenue and customer relationships.
Looking Ahead
DoorDash is expected to continue experimenting with third-party AI integrations and its own ordering assistants. The company’s experience will provide useful information about which AI features encourage additional purchases and which simply change the way existing customers access its marketplace. Partnerships involving MCP and corporate ordering could offer an early path towards more practical agent-driven transactions.
The broader AI commerce industry will need to solve issues involving trust, payments, platform access and delivery reliability before autonomous purchasing becomes a routine consumer behaviour. DoorDash’s early order numbers suggest that the transition is still at an experimental stage. Whether AI agents eventually become a major shopping channel will depend on whether they provide enough convenience and value to change how people make everyday purchasing decisions.
Frequently Asked Questions
1. How many orders has DoorDash received from AI agents?
DoorDash CEO Tony Xu said the company had received several hundred orders in total from some third-party AI agents. The company did not disclose an exact figure or a precise reporting period.
2. Which AI agents have been associated with DoorDash orders?
Third-party assistants mentioned in coverage of the earnings call include Muse, Instinct and Fo.
3. What did DoorDash CEO Tony Xu say about AI ordering?
Xu described agent-based ordering as a developing area and said customers still predominantly use DoorDash’s own application for everyday food orders.
4. What AI tools has DoorDash developed?
DoorDash has introduced Ask DoorDash, Text DoorDash, a command-line agent and an MCP integration that allows external AI applications to interact with its ordering services.
5. What is agentic commerce?
Agentic commerce refers to the use of AI agents to discover products, compare options and complete purchases on behalf of users, subject to the permissions and payment arrangements available.
6. Could DoorDash earn advertising revenue through AI agents?
Yes. CEO Tony Xu identified advertising within third-party AI assistants as a potential future opportunity. However, the company has not announced a specific advertising or revenue-sharing agreement.
7. Why is AI ordering still limited on DoorDash?
Routine food ordering is already relatively simple through DoorDash’s application. External AI agents must offer additional convenience while solving issues involving platform access, payments, trust and delivery reliability.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
