An AI revolution is reshaping the travel experiences of one billion people.

A speech by Fei-Fei Li, known as the "Godmother of AI," in 2024 propelled the concept of spatial intelligence beyond the academic circle and into the public eye. In this widely - acclaimed speech, she mentioned, "Merely seeing the world is far from enough. With spatial intelligence, AI will understand the real world."

Regarding spatial intelligence, although the industry has reached a consensus that it is one of the important evolutionary directions of artificial intelligence, there has been no clear answer as to when and how it will be implemented.

However, now a Chinese company has taken the lead in the "spatial intelligence" field. AutoNavi recently released the world's first AI - native map application: AutoNavi Map 2025.

"AutoNavi Map 2025 will transform AI from a 'conversation tool' into an 'action partner,'" said Guo Ning, the CEO of AutoNavi Map. "Different from language intelligence, spatial intelligence is the ability to perceive, reason, and act in three - dimensional space and time. It also means that our interpretation of the mission of 'connecting the real world' will further evolve to 'understanding' the real world."

01. Comprehensive AI Integration

In the past two years, AI has witnessed an explosive growth at the application level, with numerous new products emerging. However, leading applications with a large user base have generally been cautious about integrating AI, mostly limiting themselves to incremental exploration of auxiliary functions.

This is because integrating AI into a mature product is far more difficult than creating a new product from scratch. AutoNavi itself has over one billion users, and it's easy to imagine the difficulty of comprehensively integrating AI on such a large user base.

The first challenge AutoNavi needs to address is to strike a balance among different users and their diverse needs.

In terms of product experience, AutoNavi needs to ensure that the original user experience remains unaffected. Only on this basis can users gradually accept the various new AI capabilities of the product.

Moreover, AutoNavi also faces significant challenges in its technical framework.

The person in charge of AutoNavi's AI technology revealed, "In the past, we mainly optimized the original framework from the perspective of single - point optimization or local improvement, or enhanced specific links and modules. But now, we need to transform the map product itself into a more comprehensively AI - integrated form from a product perspective."

Despite numerous challenges, AutoNavi still chose to take this step, with the confidence coming from its two - decade - long accumulation.

Map is currently the largest carrier for virtualizing and digitizing the real world, and AutoNavi is one of the technology companies with the largest scale of processing spatio - temporal information such as positioning, point clouds, and vision globally.

Over the past 20 years, through continuously building a centimeter - level road network digital base covering the entire territory of China and accumulating tens of trillions of real navigation behaviors, traffic flow states, and environmental perception data, AutoNavi has established a globally scarce dynamic spatio - temporal knowledge engine.

This is not only a data reserve but also the core cognitive infrastructure for understanding complex travel systems. Every inch of road, every navigation, and every traffic jam in China has become a sample for AutoNavi's spatio - temporal training.

As Guo Ning said, AutoNavi's more than two - decade - long production of physical world data and technological accumulation provide continuous impetus for the understanding and generation of the integrated three - dimensional virtual world.

Therefore, what AutoNavi has built is not an ordinary conversational AI, but an AI travel intelligent agent. In addition to the standard Agent framework, it can also be highly aligned and adapted to AutoNavi's unique spatial data and tools.

The person in charge of AutoNavi's AI technology said, "Based on geographical data and user spatio - temporal location data, this Agent framework can understand complex spatial relationships and user needs, and efficiently execute and reflect using AutoNavi's rich travel tools to solve user problems."

Therefore, AutoNavi's comprehensive AI integration is not a short - term technological speculation but a natural evolution based on its past accumulation. From another perspective, it is also because of this AI reconstruction that AutoNavi's two - decade - long accumulation has shown greater value.

02. A Smarter AutoNavi

When comprehensive AI integration is reflected in the AutoNavi Map app, it brings about a fundamental reconstruction of the core usage chain of the product.

In AutoNavi Map 2025, users can start the core interaction mode - voice communication with AutoNavi's main intelligent agent, "Teacher Xiaogao" - by clicking the voice icon in the search bar or clicking "Conversation" at the bottom of the home page.

Whether it's customizing a multi - day cross - city self - driving tour strategy before a trip, inquiring about airport VIP lounge privileges, or adjusting and changing the travel plan in real - time during a trip, Teacher Xiaogao can accurately parse the requirements with its powerful semantic understanding ability, call sub - intelligent agents and corresponding tools after reasoning and thinking, and quickly output personalized solutions.

In addition, the full - scenario functions such as AI Navigation, AI Instant, AI Exploration, and AR Check - in in AutoNavi Map 2025 also fully cover the full - cycle needs of users from pre - trip decision - making, in - trip navigation to post - trip services.

For example, when users open the home page of AutoNavi Map and swipe down, they can see "AI Instant," which is recommending travel needs and destinations. It can accurately predict users' immediate travel needs based on the "time progression + space evolution" dual - axis sorting model and the current spatio - temporal location, and actively plan the next itinerary.

The person in charge of AutoNavi's AI products said, "AutoNavi users often have multiple needs at the same time. In the past, we only met single - search needs and did not fully explore the connections between different needs."

Therefore, when designing the "AI Instant" function, the first consideration was to reorganize user needs by "events" rather than in the original way.

For example, "business trip" is an event, which includes services such as taking a taxi, buying an air ticket, airport pick - up, and hotel reservation. "AI Instant" will first organize the supply according to events and then match the events with users' current states.

Under this logic, when users have a business trip need, AutoNavi Map will recommend flight reminders three to five days in advance, prompt users to book a car a few hours before the trip on the same day, recommend luggage check - in and security check services after arriving at the airport, and recommend lounges after passing through security. In this way, it can meet users' needs throughout the entire travel lifecycle.

The person in charge of AutoNavi's AI products said that users' travel behaviors are highly time - sensitive and space - sensitive. If there is no response at critical spatio - temporal nodes, users' needs may change in a short period or over a short distance.

Now, AutoNavi hopes to meet these user needs with its spatial intelligence capabilities.

03. The Future of Spatial Intelligence

Currently, spatial intelligence is one of the most promising AI tracks, and AutoNavi has thus embarked on a bright path.

However, this path is not smooth. AutoNavi's early entry means that it will shoulder more responsibilities to push forward spatial intelligence, a long - term track that is still in the "experimental stage," towards the future.

AutoNavi's evolution in spatial intelligence can be regarded as a "genetic recombination" of the map product - it enables a software to actively perceive, think, and predict rather than passively respond, and ultimately drive personalized decision - making.

These are also the core values of spatial intelligence: achieving a closed - loop ability of perception, reasoning, and action in space and time. Therefore, the implementation of spatial intelligence cannot rely solely on concept and single - point technological breakthroughs but also needs to be deeply rooted in scenario exploration and data accumulation.

Of course, for AutoNavi, the realization of spatial intelligence will not happen overnight; it will be a long - term project. Especially from the user's perspective, spatial intelligence is still difficult to be perceived continuously at present, but it will surely become an indispensable infrastructure layer in people's digital lives in the future.

In the future, AutoNavi's spatial intelligence may become a "router" connecting the physical world and the digital world, bringing a broader imagination far beyond the current scenarios.

For example, in the field of local life, spatial intelligence is expected to reshape the consumer decision - making chain, driving users' needs to shift from "how to get there" to "how to have fun." Imagine that in the future, the system can recommend destinations based on users' current emotional states and physical fatigue levels, recommend the best shooting locations and time periods based on historical culture and social media hotspots, and connect the closed - loop of planning, navigation, sharing, and consumption conversion. This will undoubtedly create a new entry point for local life.

Even grander prospects are hidden in emerging industries such as low - altitude economy and embodied intelligence. In the future, spatial intelligence may become the invisible framework of smart cities and various robots.

Back to the present, AutoNavi, which has shown initial results and clear user value, is like the "first - phase project" of the grand spatial intelligence project, allowing us to see the imagination behind its perseverance.

This year will undoubtedly be a year when the "depth of scenarios" of AI is more important than the "speed of concepts." The stage of spatial intelligence has just begun, and the value returns generated by AutoNavi's long - term and firm investment in technology and scenarios have quietly accumulated in billions of trips and searches.