Major internet companies are accelerating the deployment of AI products and applications. Following the introduction of AI features for shopping, food delivery, and hotel bookings by several domestic tech giants, another essential daily scenario has now integrated AI technology.
After initiating public testing of its AI travel assistant in September 2025, DiDi Global Inc. recently launched an upgraded v1.0 version. This marks the industry's first AI-driven ride-hailing service, making it available to all registered users. By using the "AI Ride-Hailing" feature in the DiDi app, customers can simply state their request in one sentence. Whether it's a personalized vehicle requirement or a vague description of their situation, the AI assistant, XiaoDi, can help users find a vehicle that meets their needs.
Tests conducted after updating the app show that DiDi's new AI ride-hailing function can accurately match various user needs with just a single voice command. These include complex scenarios such as requesting a car that minimizes motion sickness, specifying multiple stops, needing a vehicle suitable for pregnant women, or preferring newer car models.
Analysts point out that large AI models are now entering a phase of practical implementation, with industries moving toward refined operations and commercial verification. Companies that possess specialized industry data and a deep understanding of specific scenarios have a competitive edge in developing tailored AI solutions. Such firms can integrate advanced AI technology with their core business operations to achieve differentiated market positioning.
DiDi has officially rolled out its AI ride-hailing service by embedding an "AI Ride-Hailing" module just below the commonly used input field in its app. This allows for highly personalized ride bookings. It represents another significant case of an internet company accelerating the practical application of AI since 2026.
In practical tests, after stating ride-hailing needs—whether simple or complex—the AI assistant XiaoDi initiates a matching process based on the user's command. It then presents three vehicle options, each accompanied by tags such as "new car," "spacious interior," or "smooth driving" to assist users in making an informed selection.
When testing more complex scenarios, it was observed that the function can achieve precise vehicle matching based on detailed user requirements. For example, when a voice command was given: "I need to go from Chaoyang Park to Terminal 3 of Beijing Capital Airport with elderly people and children. The car must have fresh, clean air and should be spacious if possible," the AI assistant processed the request within three seconds. It automatically matched vehicles tagged with "smooth driving," "spacious interior," "no odors," and "new car." Within ten seconds, it successfully identified the nearest available vehicle meeting these criteria. Users only need to tap "Confirm Ride" and wait for the car to arrive at their location.
It is understood that the AI assistant currently supports over 90 service tags, including "fresh air," "large trunk," and "smooth driving," covering complex travel scenarios such as traveling with family or business receptions.
Beyond one-command ride selection, the AI assistant also offers multiple other functions, such as "nearby stores," "schedule a ride," "order inquiry," and "combined travel planning." For instance, when asked to "help plan a trip to Beijing Daxing Airport," the AI generated a combined route using subway and ride-hailing options, providing precise details on travel time, cost, and distance.
As AI models evolve rapidly, many internet firms are speeding up the rollout of vertical AI applications, with user lifestyle scenarios becoming a key competitive area. In the ride-hailing sector, DiDi, as the first platform to introduce an AI-powered service, is driving further upgrades in industry standards.
Test results indicate that DiDi's current AI ride-hailing service can address the vast majority of user needs. The AI assistant translates user speech into executable platform tags. For example, mentions of "not feeling well" or "motion sickness" trigger tags like "smooth driving" and "gasoline car." References to "pregnant" activate tags such as "smooth driving" and "spacious interior." These are combined with real-time traffic, time, vehicle location, and driver status to quickly filter available options in the dispatch pool, finally presenting candidate choices for user confirmation.
Notably, even when a perfect match is not immediately available, DiDi's AI assistant can prioritize complex requirements—addressing core needs first and offering the most feasible available option.
The more personalized the request, the greater the matching challenge. The AI must not only understand natural language but also operate within dynamic constraints like changing traffic conditions and real-time supply and demand. This relies not just on model capability but also on long-accumulated systemic strengths, where scale, service control, and operational experience play decisive roles.
After more than a decade of operation, DiDi has accumulated a vast repository of authentic user reviews and service tags. These underlying capabilities enable the platform to accurately interpret specific instructions—such as which car offers a fresher environment or which driver provides a smoother ride—allowing the AI to precisely meet individualized customer demands.
Industry observers note that the introduction of DiDi's AI ride-hailing service will shift the industry focus from "scale competition" to "value competition." Key impacts include: First, service certainty is enhanced as AI converts vague requests into precise service tags, putting an end to the uncertainty of random vehicle assignment and improving user experience. Second, drivers offering superior service will receive more orders, creating a positive cycle of "better service leading to higher income and further service improvements." Third, technological barriers become more pronounced; advantages in scale, standardized service, and scenario expertise form a moat that is difficult to replicate in the short term.
According to some analysts, AI serves as a core driver for stimulating new consumption growth, aligning well with policy directions. It is deeply penetrating consumer applications through "scenario power." Driven by both policy support and market dynamics, AI is becoming a central engine for activating new consumption growth points and expanding domestic demand.
Other analysts add that enterprises with specialized industry data and scenario comprehension are better positioned to develop vertical large models and AI agents. By deeply integrating advanced AI technology with their business operations, these companies can achieve a distinct competitive advantage.
Comments