Amap, Alibaba Group’s map navigation unit, recently announced its full AI transformation with the launch of "Amap 2025" and introduced the world’s first map-based AI-native agent.
Amap CEO Ning Guo said the company is evolving from a traditional navigation tool to a "spatial intelligence agent," enabling maps to not only depict roads but also interpret dynamic traffic, geographic features, contextual scenes, and human intentions—achieving a "see—reason—act" capability.
By combining 20 years of accumulated data with modern large models, Amap has advanced its maps from "static base maps + passive planning" to "dynamic cognition + proactive decision-making," building an "action engine" for the physical world.
China’s complex and diverse traffic environment offers Amap AI a unique real-world testing ground. Over two decades, Amap has built massive spatiotemporal datasets on roads, buildings, terrain, points of interest, vehicle flows, and pedestrian flows, supporting a closed-loop training system that continuously iterates between virtual simulations and real-world validation.
For safety alerts, Amap provides both near-range and far-range warnings, generating over 60 million daily notifications. In 2024, during a highway collapse incident, the Amap system quickly detected anomalies and issued precise alerts to vehicles behind, demonstrating the practical capabilities of spatial intelligence.
Amap has embodied this spatial intelligence in the conversational agent "Xiao Gao," designed to deliver full-cycle services spanning "understand—plan—reflect—act." Users can request cross-city self-driving routes, temporary detours, and service bookings, with Xiao Gao generating personalized solutions via spatial reasoning and tool integration.
This year, features like "AI Instant" and "AI Explore" were launched to combine real-time responses in high-frequency scenarios with deep planning capabilities. The system integrates pre-trip, in-trip, and post-trip processes, optimizing routes, avoiding congestion, and arranging rest stops, creating a "living map" experience.
Ning Guo emphasized that Amap’s AI transformation is not a trend-chasing move but a natural evolution driven by technology and demand. He described spatial intelligence as part of Amap’s DNA and the vital conduit connecting the digital and physical worlds.
With long-term data accumulation, algorithmic advancements, and ecosystem openness, Amap aims to transform AI from a "conversational tool" into an "action partner," providing users with reliable, safe, and intelligent mobility services while empowering industries such as autonomous driving, drone logistics, and robotic delivery.
Looking ahead, spatial intelligence could become the digital core for urban governance, restructuring travel efficiency and enabling all smart devices to understand and interact with the physical world.
