The rapid rise of artificial intelligence agents has sparked widespread industry chatter about an impending autonomous takeover of commerce, where software assistants handle everything from shopping lists to meal ordering on behalf of human users. However, according to delivery giant DoorDash, that dystopian or utopian future—depending on how one views autonomous digital labor—is still quite far off on the horizon.
During its recent third-quarter 2026 earnings call, DoorDash revealed that despite the intense hype surrounding agentic AI, it has only recorded "several hundred" total orders originating from third-party autonomous agents like Muse, Instinct, Fo, and various other emerging platforms. To put that minuscule figure into stark perspective, the company disclosed in its 2025 operational report that it routinely completes an average of roughly 3 million deliveries every single day.
The vast chasm between millions of daily human-driven orders and a few hundred agent-driven transactions highlights just how nascent the agentic commerce ecosystem remains. While tech enthusiasts and software developers envision a world where autonomous agents seamlessly manage our daily logistics, the practical reality on the ground tells a much more measured story.
Addressing the trend during a discussion with financial analysts, DoorDash CEO Tony Xu offered a balanced perspective on the intersection of artificial intelligence and food delivery. He acknowledged the rapid pace of technological innovation while cautioning against overestimating how quickly consumer habits actually change.
"I would say that this [using agents to order] is something that is still a very dynamic area," Xu told analysts during the earnings call. "And my take on all things technology is you never want to fight gravity. But at the same time, you don’t want to be delusional."
Xu elaborated on what that philosophy means for DoorDash’s strategic posture, pointing directly to the current volume numbers being generated by external artificial intelligence platforms.
"What I mean by that is if you look at where AI agents are today, we’re seeing maybe several hundred orders in total from some of these third-party agents," he noted, grounding the high-flying tech discourse in hard commercial metrics.
Despite the low current volume, Xu emphasized that the vast majority of consumers are continuing to place their food and retail orders through traditional pathways, most notably the primary DoorDash mobile application. Yet, the company is far from ignoring the broader shift toward conversational and agentic computing. DoorDash is actively seeking to understand how customers might leverage automated assistants in the future, aiming to cultivate collaborative, healthy relationships with the creators of third-party AI agents rather than shutting them out entirely. Looking further down the road, Xu suggested that as agentic platforms mature and capture more consumer mindshare, there could emerge lucrative opportunities for targeted advertising and brand placement within these autonomous assistant ecosystems.
DoorDash’s cautious optimism regarding external agents matches its own aggressive internal experimentation with conversational commerce and automated ordering interfaces. Over the past several months, the delivery titan has rolled out a suite of proprietary tools designed to test how consumers respond to non-traditional ordering methods.
In September, DoorDash launched a specialized text-to-order AI agent that allows users to initiate and complete deliveries simply by sending text messages. That followed a July rollout of a command-line interface (CLI) agent tailored for tech-savvy users who prefer interacting via text-based terminal commands. Earlier in June, the company introduced a sophisticated chatbot named Ask DoorDash, which enables customers to place orders using natural language prompts and even photos of food items they wish to replicate or acquire. Additionally, DoorDash has released an ordering connector protocol, known as an MCP, designed to facilitate seamless ordering directly from other third-party chatbots, primarily targeting the corporate ordering segment where administrative friction is high.
While DoorDash is working to position itself wherever the customer might be—whether inside a traditional app, a text thread, or a chatbot window—the broader digital commerce landscape is experiencing a sharp strategic division over how to handle autonomous AI agents. Retailers and platform operators are currently locked in a high-stakes debate regarding whether these software programs should be granted open access to company websites, databases, and product inventories.
The approach varies wildly among retail heavyweights. On one side of the spectrum, e-commerce titan Amazon has taken a restrictive stance, actively blocking personal AI agents from scraping its platform or executing automated purchases on behalf of users. Amazon’s defensive posture reflects long-standing concerns regarding site security, server loads, scraping, and the preservation of direct consumer relationships and ad revenue.
Conversely, big-box rival Walmart has taken the opposite path, rolling out the welcome mat for third-party AI agents. By intentionally opening its digital doors to automated assistants, Walmart hopes to capture emerging revenue streams and siphon off market share from competitors like Amazon by making its catalog as accessible as possible to whatever software a consumer happens to be using.
As major retailers chart opposing courses and actual consumer transaction volumes via third-party agents remain in the hundreds rather than the millions, the trillion-dollar question for the tech industry is how quickly agentic commerce will cross the chasm from experimental novelty to mainstream utility. For now, companies like DoorDash are keeping a watchful eye on the horizon, preparing for a potential wave of automation while recognizing that the average hungry customer is still tapping away on a smartphone screen the old-fashioned way.
