At the "AI Integration, Rejuvenating Consumption – Quality Ignites a New 'AI + Consumption' Ecosystem" forum during the 2026 World Artificial Intelligence Conference, Liu Jiaqi, Head of AI Applications at Didi Group, delivered a keynote speech titled "AI Reconstructs New Mobility Experiences, Driving Service Consumption Quality and Volume Growth."
It was noted that on the Didi platform, user mobility needs extend far beyond commuting and visiting family or friends, encompassing a significant volume of consumption-oriented travel for shopping, dining, entertainment, and tourism. The consumption linkage effect driven by mobility is remarkably pronounced. Didi's research indicates that, on average, every $1 spent on ride-hailing can stimulate several times that amount in other offline expenditures. However, for a long time, some passengers' personalized travel requirements have not been precisely met. From Didi's perspective, mobility is not merely about moving from point A to point B; it acts as a crucial hub connecting consumers with offline consumption scenarios. When more personalized travel demands are unlocked by AI, the efficiency and potential of offline consumption are consequently activated.
The emergence of AI has opened up substantial "personalization" space for mobility. The AI assistant, Xiao Di, already possesses over 90 tags to cater to diverse scenario-based travel needs. Simultaneously, it pragmatically aids decision-making by balancing experience and efficiency. AI translates users' vague requirements into definitive transactions. Leveraging Didi's leading supply resources and dispatch capabilities, the service evolves from "finding a car" to "finding the right car," which is a concrete manifestation of "good service" in the AI era.
Key Advancements in AI Matching
Liu Jiaqi stated that one of the core breakthroughs of AI Xiao Di is a personalized ride-hailing matching system, coupled with a driver-side tag system for identification and distribution. This system has been used to optimize dispatch algorithms, promoting a two-way balance between supply and demand: on one hand, it satisfies passengers' personalized, multi-scenario ride-hailing needs; on the other, it allows drivers with better vehicle conditions and superior service to secure more ride opportunities. Passenger satisfaction, driver-side service quality, and increased travel and consumption demand form a positive feedback loop here. On the technical front, Xiao Di employs a decentralized multi-agent architecture, which has improved satisfaction in core ride-hailing scenarios by nearly 10 percentage points.
Beyond personalized ride-hailing, Xiao Di has progressively rolled out features such as "Search Along the Route," "Multi-Destination Planning," "Surrounding Recommendations," and "Complex Address Pick-up/Drop-off Point Recommendations."
Future Outlook for AI Products
Looking ahead at the development of AI products, Liu Jiaqi believes transactional Agents may become new infrastructure for service consumption, and the fulfillment of personalized needs will evolve into a universally accessible service. Under the guidance of relevant authorities, DiDi Global Inc. is collaborating with AI industry partners to jointly establish credible standards for industry development. This effort aims to promote the healthy growth of AI-powered mobility services and inject sustained momentum into enhancing the quality and volume of service consumption.
Comments