What Uber reportedly counted
TechCrunch reports that Uber recorded four rides booked through AI agents during its busiest week in October. The report, republished by Yahoo, also cites 500 agent-placed Uber Eats orders during the Halloween week.
The company handled around 320 million trips and orders across Uber and Uber Eats in that period, according to the report. That combined total is useful scale context, but it is not a rides-only denominator and should not be used to invent a precise agent share of ride bookings.
A task must justify the extra step
The report attributes the figures to Uber’s earnings call and describes management as still positive about agents becoming another route to the platform. It also quotes the company’s view that its existing app already makes common bookings quick.
Franklin’s analysis is that an agent needs to remove meaningful work, rather than simply add another interface. A routine repeat order may require little planning. A complicated request involving constraints, comparison and coordination poses a different test. Low use in the reported transaction category does not establish that every kind of agent task lacks value.
What would make the adoption picture clearer
These figures describe an early period and the bookings the company classified as agent-originated. They do not count every consumer who consulted an assistant before completing an order manually.
Useful follow-up evidence would separate discovery, recommendation and completed transactions; it would also track failed attempts and repeat usage. That would show whether the bottleneck is user demand, integrations or the reliability of completing a booking. Until those distinctions are available, the defensible conclusion is specific: the reported number of directly agent-booked rides was very small.