Almost no one uses AI agents to book their Uber, company says

Uber disclosed during its third-quarter earnings call on Wednesday that almost no customers are using autonomous software tools to book transit or request meal delivery. Chief Financial Officer Balaji Krishnamurthy reported that during the week of Halloween in October, which was the platform's busiest week of the quarter, customers completed only four ride bookings through an automated software agent. During that identical week, the overall platform processed roughly 320 million completed trips and food orders across the core ride-hailing network and Uber Eats, while AI-initiated requests on Uber Eats totaled just 500 meal deliveries. According to TechCrunch's report, chief executive officer Dara Khosrowshahi stated that traction remains very early and limited across consumer assistants, though leadership still anticipates broader user adoption as foundational software ecosystems mature.

This breakdown examines the transaction records shared with Wall Street analysts, the friction points limiting automated workflow adoption across consumer services, and how competitive platforms like DoorDash measure up against similar software integrations. It is intended for software engineers, product managers, and enterprise planners evaluating whether autonomous consumer agents can replace native mobile application interfaces for high-frequency, on-demand physical logistics.

Almost no one uses AI agents to book their: what actually changed

Uber provided concrete operational metrics revealing that personal software assistants accounted for a statistically negligible share of transactions during peak operational volume. Chief Financial Officer Balaji Krishnamurthy confirmed to financial analysts that out of 320 million total trips and delivery orders handled during the final week of October, exactly four rides were reserved using an AI intermediary. Meal orders initiated through external automated assistants on Uber Eats reached 500 transactions over the same seven-day window, demonstrating that food delivery saw slightly higher activity than passenger transport but still represented an infinitesimal fraction of overall volume.

The low utilization figures arrived despite heavy capital investments and surging consumer visibility across the broader generative assistant ecosystem. Startups such as Instinct, which launched last year, secured funding rounds reaching a 10 billion dollar private valuation, while Meta stated that its Muse software reached 10 million users and climbed to the top position on Apple's U.S. App Store last month. Tech giants including OpenAI, Anthropic, Google, and SpaceXAI through its Grok platform have all released automated task assistants, yet the disclosed audit from Uber's management indicates that consumer interactions remain almost entirely contained within standard user interfaces rather than delegating tasks to external agents.

Almost no one uses AI agents to book their: how it works

Autonomous agent booking relies on conversational software tools translating user instructions into programmatic transactions across external merchant interfaces and transport networks. When a consumer instructs an assistant to arrange a trip or place a takeout order, the agent must interpret destination criteria, parse pickup locations, navigate variable vehicle tiers, handle payment tokens, and verify schedule constraints without direct manual input on the ride-hailing screen. For restaurant deliveries through Uber Eats, the agent must parse specific dietary modifications, locate menu selections, verify merchant availability, and submit the checkout sequence directly into the service's transaction intake pipeline.

Uber leadership highlighted that native smartphone interfaces already minimize transactional friction to a degree that automated agents struggle to improve upon. Dara Khosrowshahi pointed out during the earnings presentation that an ordinary passenger requires an average of just 50 seconds to complete a ride reservation directly inside the native Uber smartphone application. Similarly, repeating a preferred meal order inside Uber Eats requires an average duration of only 90 seconds, indicating that routing complex natural-language queries through multi-step third-party reasoning models offers little tangible time savings to everyday consumers.

Almost no one uses AI agents to book their: why it matters

The minimal booking volume underscores a wide divergence between software industry valuations and verifiable consumer behavior within autonomous task platforms. Venture capital and technology providers have spent months marketing automated assistants as an imminent replacement for direct web navigation and individual mobile apps, yet transaction volumes on public logistics networks demonstrate that consumers still prefer direct operational control. DoorDash reported comparable consumer behavior during its own earnings release, noting that it had received only a few hundred food orders via autonomous agents, with the overwhelming majority of diners continuing to order through DoorDash's native mobile interface.

Despite current transaction figures totaling four rides and 500 meal orders, Uber executive management framed external automated tools as an incremental distribution channel rather than an existential business threat. Khosrowshahi described autonomous systems as an additional mechanism to access the underlying transport fleet and expressed optimism that automated integrations will eventually create a net positive impact on ecosystem demand. The platform's perspective treats conversational interfaces as auxiliary digital endpoints that must ultimately route physical dispatch and fulfillment through established logistics infrastructure.

Almost no one uses AI agents to book their: who is affected

The immediate impact falls upon autonomous software developers who must prove that conversational task execution can replace high-speed mobile consumer software. Emerging agent platforms such as Instinct, established offerings like Meta's Muse, and enterprise models from Google, OpenAI, Anthropic, and SpaceXAI must justify complex backend tool-calling pipelines against existing single-tap user flows. When automated handoffs fail to beat a 50-second app interaction or a 90-second repeat meal purchase, consumer incentives to route payments and credentials through third-party conversational assistants remain extremely weak.

For transit riders and restaurant customers, direct mobile software remains the dominant, most dependable channel for everyday service dispatch. Consumers continue opening dedicated smartphone apps to track vehicle locations, confirm driver identities, verify real-time surge pricing, and ensure delivery accuracy instead of relying on third-party conversational agents. On-demand logistics operators like Uber and DoorDash maintain their direct relationship with end users, avoiding immediate disintermediation while assessing technical protocols alongside software developers.

Almost no one uses AI agents to book their: what to watch

Industry attention now shifts to commercial negotiations between digital assistant providers and transport infrastructure operators. Khosrowshahi noted that approximately eight prominent personal AI assistants are currently active across consumer markets, adding that he expects that group to expand beyond ten providers over time. Furthermore, the chief executive confirmed that every major AI system developer is currently in technical discussions with Uber to establish formal platform integrations, which could standardize how conversational assistants query pricing and trigger ride dispatches in future releases.

Monitoring whether quarterly numbers move beyond hundreds of experimental requests will determine if autonomous assistants gain genuine operational utility. Analysts will watch subsequent earnings disclosures from Uber, DoorDash, and peer delivery networks to evaluate whether deeper operating system integrations or improved conversational accuracy can compress booking times below native mobile benchmarks. Until automated task workflows demonstrate obvious speed advantages over dedicated smartphone apps, agentic consumer bookings will remain a marginal slice of overall digital commerce.