DiDi's AI travel assistant, which began public testing last September, has now officially launched its XiaoDi v1.0 version. The service currently supports more than 90 service tags, such as "fresh air," "spacious trunk," and "smooth driving," covering a wide range of complex travel scenarios including assisting the elderly and children, as well as business receptions. Users can now update the DiDi app to the latest version, click on "AI Ride-Hailing" on the homepage to experience XiaoDi, and receive corresponding incentives for trying the feature.
According to the introduction, XiaoDi translates user requests into executable platform tags. For example, if a user mentions "feeling unwell" or "motion sickness," the system will activate tags like "smooth driving" and "gasoline vehicle." If "pregnant" is mentioned, tags such as "smooth driving" and "spacious interior" will be triggered. Combining these with real-time traffic conditions, time, vehicle location, and driver status, XiaoDi quickly filters options from the dispatch pool and presents candidate choices to the user for confirmation. If no perfect match is available, XiaoDi can prioritize complex needs by first addressing core requirements and offering a practical "better solution" for the moment.
Beyond selecting a ride with a simple voice command, XiaoDi is expanding into a more comprehensive travel assistant. It also supports features like searching for nearby locations and booking a ride with one click, recommending transfer options for long-distance trips, checking order details, and scheduling rides in advance.
For AI to not only understand human language but also match rides under constantly changing constraints like complex road conditions and real-time supply and demand, it relies not only on model capabilities but also on DiDi's long-accumulated systemic strengths. The first key factor is scale. With a sufficiently large supply of drivers, the platform can afford to break down user demands into finer details. Otherwise, the more specific the request, the harder it becomes to book a ride. The second factor is service control. Under its self-operated model, DiDi directly serves both drivers and passengers, leading to more standardized and controllable service quality. The third factor is data accumulation. Over a decade of operation, the platform has amassed vast amounts of authentic reviews and tag data, which determine its ability to accurately answer questions like "which car has fresher air?" or "which driver offers a smoother ride?" This data is crucial for evolving AI dispatch from merely "guessing what you like" to truly "understanding what you need."
Comments