Industry First! Didi Makes a Big Move! Has Ride-hailing Entered the "One-Word Era"?

Major internet companies are accelerating the deployment of AI products and applications!

Following the launch of AI shopping, food delivery, hotel booking, and other functions by several domestic internet giants, another important life scenario is now embracing AI applications.

Since September 2025, when Didi launched its AI travel assistant beta, it has recently upgraded to version 1.0. This is also the industry’s first AI ride-hailing service, meaning all registered users can access it. Users can experience “AI ride-hailing” within the Didi app—just one sentence, whether they have personalized vehicle requests or vague descriptions of their own condition, AI Xiaodi can help find vehicles that meet their needs.

After updating the latest version of the app and testing, securities industry reporter found that Didi’s new AI ride-hailing feature can accurately match multiple user needs with just one sentence: for example, avoiding motion sickness, multiple stops, pregnant women, new or old vehicle models, and other complex scenarios.

Several analysts pointed out that AI large models are currently entering a practical implementation phase, with the industry moving into a new stage of refined operations and commercial validation. Companies with data barriers in specific industries and scene understanding capabilities have advantages in developing industry-specific AI agents, enabling deep integration of large model technology with their own businesses to achieve differentiated competition.

Reporter Test: Personalized ride-hailing with just one sentence

Recently, Didi officially launched its AI ride-hailing service by adding an “AI ride-hailing” module below the commonly used input box in the Didi app, enabling personalized ride services. This is another significant case of internet companies accelerating AI application deployment since 2026.

After updating the Didi app, securities industry reporter tested the feature by verbally stating their ride needs, whether simple or complex. AI Xiaodi would initiate demand matching based on user instructions, listing three options for selection. Different vehicle options also include tags like “new car,” “spacious interior,” and “smooth driving,” helping users reference and choose.

In testing complex travel scenarios, the reporter found that the feature could precisely match vehicles based on detailed user needs.

For example, using voice input, the reporter told the AI assistant, “Now going from Chaoyang Park to Terminal 3 of Capital Airport, I’m with elderly and children, and I need fresh air and comfort, preferably spacious,” and after three seconds, the AI assistant automatically matched vehicles with tags like “smooth driving,” “spacious interior,” “no odor,” and “new car,” successfully finding the closest suitable model within ten seconds. After clicking “Confirm Ride,” the vehicle would arrive at the user’s location.

Currently, the AI assistant has over 90 service tags, including air freshness, large trunk, smooth driving, covering more complex travel scenarios like elderly and children care, business reception, and more.

The reporter also observed that besides one-sentence vehicle selection, the AI assistant offers other functions such as “Nearby stores,” “Book a ride,” “Order inquiry,” and “Multi-modal travel.” For example, when asked to “help plan a trip to Daxing Airport,” the AI analyzed and generated a combined route using subway and ride-hailing, providing precise travel time, cost, and distance estimates.

Ride-hailing industry shifts from “scale competition” to “value competition”

In fact, as AI large models rapidly iterate, many internet companies are accelerating the deployment of vertical AI applications, with user life scenarios becoming key battlegrounds, and competition intensifying. Taking the ride-hailing industry as an example, Didi, as the first to launch AI ride-hailing services, has driven further upgrades in transportation services.

From the test results, Didi’s current AI ride-hailing service can meet most user needs.

For example, the AI assistant translates user speech into platform-executable tags. If it detects “feeling unwell” or “motion sickness,” it activates tags like “smooth driving” and “gas vehicle.” If it recognizes “pregnant woman,” it activates “smooth driving” and “spacious interior.” Combining real-time traffic, time, vehicle location, and driver status, it quickly filters options in the dispatch pool and presents candidate cards for user confirmation.

Notably, if no perfect match exists, Didi’s AI assistant can prioritize complex needs: first satisfying core requirements and pragmatically providing the “best current solution.”

In reality, the more personalized the request, the harder the matching. AI must understand human language and match under complex road conditions, real-time supply and demand, and other rapidly changing constraints. Behind this, it relies not only on model capabilities but also on long-term system accumulation. Scale effects, service control, and experience sedimentation are three critical factors.

According to securities industry sources, Didi’s decade-plus operational experience and real user evaluations and tags enable it to accurately match specific commands like “which car is fresher” or “which driver is more steady,” allowing precise AI dispatch to meet personalized customer needs.

Industry analysts believe that the launch of Didi’s AI ride-hailing will shift the industry from “scale competition” to “value competition,” with three main impacts: first, service certainty will improve as AI transforms vague needs into precise service tags, ending “blind box” ride-hailing and enhancing user experience; second, high-quality drivers will receive more orders, creating a positive cycle of “good service—high income—better service”; third, technological barriers will become more prominent, with scale capacity, standardized services, and scene accumulation forming a moat that is difficult to replicate in the short term.

Guojin Securities believes that AI, as a core driver of new consumption growth, aligns well with policy guidance and is deeply penetrating consumer applications through “scene capability.” Under the dual drive of policy and market, AI’s deep integration into all consumer scenarios has become a key engine for activating new consumption growth and expanding domestic demand.

BOC Securities notes that companies with data barriers in specific industries and scene understanding capabilities have advantages in developing industry-specific large models and AI agents, enabling deep integration of large model technology with their own businesses to achieve differentiated competition.