DiDi(US:DIDIY) Xin Lang Cai Jing·2026-03-17 11:22Core Insights - Didi's AI travel assistant, Xiaodi v1.0, has officially launched after a public beta in September last year, supporting over 90 service tags for complex travel scenarios [1][3] - Users can upgrade the Didi app to access Xiaodi, which translates user requests into actionable tags and prioritizes core needs when perfect matches are unavailable [1][3] Group 1: Features and Functionality - Xiaodi supports various service tags such as air freshness, spacious trunk, and smooth driving, catering to diverse user needs including elderly care and business reception [1][3] - The assistant can identify user needs like "feeling unwell" or "pregnant" and adjust the ride accordingly by selecting appropriate tags and considering real-time conditions [1][3] - In addition to ride selection, Xiaodi offers features like nearby location queries, one-click ride requests, long-distance transfer recommendations, order inquiries, and ride reservations [1][3] Group 2: System Capabilities - Didi's AI relies on not just model capabilities but also on long-term accumulated system capabilities, including scale effects, service control, and data accumulation [2][4] - Sufficient supply allows for detailed demand segmentation, making it easier to fulfill specific ride requests [2][4] - Didi's self-operated model ensures standardized service quality for drivers and vehicles, enhancing overall service reliability [2][4] - The platform's decade-long operational data provides accurate evaluations and tag data, enabling a shift from "guessing" to "understanding user needs" in AI dispatch [2][4]