Tencent Maps Opens Up Full Agent Capability Landscape and Releases LBSwork

Tencent Maps wants to bring all Agents into its circle of friends. Letter List App Download the client for exclusive early access. November 02, 2026 15:53:29 From Beijing Updated November 02, 2026 17:59:44

Tencent Maps' answer is "ecosystem."

Its goal is to prioritize helping ecosystem partners improve map-related capabilities, serving as the map foundation for the AI era and forming a thriving LBS ecosystem.

Several of Tencent's core AI products—Yuanbao, Xiaowei, and WorkBuddy—have long been deeply integrated with Tencent Maps.

When users ask questions, look up information, orchestrate tasks, and set goals for Agents in these AI apps, anything involving geographic location services can increasingly be supported at the underlying level by Tencent Maps.

But a vibrant ecosystem shouldn't consist only of "family." Tencent Maps is opening its arms to third-party companies across various fields. In addition to hardware devices or system-level AI on phones such as Rokid AI glasses, Guangfan AI earbuds, and OPPO's Breeno assistant, this also includes ByteDance's Doubao.

On the morning of October 28, Letter AI and other media held a discussion with Tencent Maps management. Zhong Xiangping, Vice President of Tencent Group and President of Tencent Smart Mobility, said that Tencent Maps will not close its doors to any partner and will pursue open cooperation.

Caption: Zhong Xiangping, Vice President of Tencent Group and President of Tencent Smart Mobility

Zhang Zhidong, Vice President of Tencent Maps, also said, "We have provided all our capabilities; we have a fully open attitude."

The signal Tencent Maps hopes to send is clear: it will adhere to its positioning as map infrastructure, not proactively set limits, not overshadow others, and serve internal and external partners well.

Tencent Maps is stepping into the Agent era together with Yuanbao, Doubao, and others.

In the past, users opened various map apps, or internet services with embedded maps, to obtain locations and routes. Now, in addition to information-based needs, question-and-answer, service-oriented, and request-based needs have emerged. Beyond "checking locations" and "looking up routes," users also want to "find guides" and "find services."

The paradigm of interaction between people and maps is being rewritten. Map products must adapt to this change.

On October 31, at the 2026 Tencent Global Digital Ecosystem Summit, Tencent Maps released and fully opened its Agent capability panorama, transforming maps, locations, travel, and local services into capabilities that AI Agents can invoke, enabling AI to provide users with a map they can open, use to look up routes, continue asking questions about, and that offers a native experience.

The goal of this panorama is to "open maps to every AI and developer," with the core being "Agent-friendly map capabilities."

MCP Server points to the protocol layer, packaging map tools into standardized interfaces. It addresses "how maps connect with third-party tools" and serves as the transmission pipeline connecting Agents with map capabilities.

Agent Skill points to the scenario layer, turning map, development, travel, and other capabilities into reusable skill modules. It addresses "what skills Agents need in different scenarios."

The three form a "capability-pipeline-skill" collaborative relationship: AI services are responsible for understanding needs and providing decisions, MCP Server is responsible for standardized transmission and orchestration, and Agent Skill is responsible for turning decisions into concrete actions.

Together they support upper-layer Agents such as map cards, search, and guides, ultimately serving industry scenarios including AI coding/office work, smart terminals, phone/vehicle assistants, and industry SaaS.

Zhang Zhidong said that Tencent Maps is not just making maps callable by AI, but enabling them to understand users, organize solutions, mobilize services, and proactively appear in appropriate scenarios, becoming the location and travel service foundation for Agents connecting to the real world.

At present, these capabilities have already been implemented in numerous products and scenarios.

For example, when a WeChat user opens Xiaowei and asks, "Where is Shenzhen Sea World?" Xiaowei will invoke a Tencent Maps card for display. If the user wants to learn more, they can jump to the corresponding mini program; the intent, presentation, search, and query throughout the entire process are all completed jointly by Tencent Maps and Xiaowei.

After the Agent capabilities were enhanced, feedback from users followed.

Zhang Zhidong revealed that after Xiaowei integrated Tencent Maps, the proportion of user requests related to places and locations was initially very small, but it is now growing rapidly.

At the same time, it and the Tencent Maps mini program form a mutually reinforcing, spiraling relationship, and have not cannibalized mini program usage.

Beyond consumer users, enterprise users large and small also have new expectations for maps.

For example, site selection in the retail industry is time-consuming and labor-intensive. Enterprise technical staff need to first read documentation, understand parameters, process data, and build pages before they can generate a site selection map. With the arrival of the AI era, one-click generation via Agents has naturally become an enterprise expectation.

Similar situations are also very common in SaaS, travel, marketing, and other fields. Enterprises want to focus more energy on the business itself rather than on how to develop maps and invoke processing interfaces.

Around this need, Tencent Maps released a new product, LBSwork, which can be called the "map version of WorkBuddy."

LBSwork is a desktop AI map workbench. Based on the same Harness framework as WorkBuddy, it deeply integrates Tencent Maps' spatiotemporal dynamic large model, consolidates Tencent Location Services and location big data capabilities as well as vertical industry knowledge, and can provide one-stop "AI + map" services for customers in retail, culture and tourism, travel, and other industries.

The "Expert Think Tank," which integrates multiple experts, is the core feature of LBS.

This feature organizes professional knowledge and tools around different map scenarios, providing multiple experts for users to choose from, including map visualization experts, daily travel experts, marketing insight experts, retail site selection experts, ride-hailing experts, crowd heat experts, logistics fulfillment experts, and more.

These experts each have their own focus: some are oriented toward map visualization and data insights, while others provide LBS services for vertical scenarios. It is understood that LBSwork will subsequently add more experts to cover more scenarios.

Still taking retail site selection as an example: in the past, site selection relied on legwork and reading reports. Now, with the assistance of LBSwork, business personnel only need to have a natural conversation to compare candidate locations, sort out advantages and risks, and obtain reviewable site selection recommendations based on the business format, customer groups, and site selection criteria.

Yu Xiaodong, general manager of Beijing Leshi Weiguang Network Technology Co., Ltd., said that guesswork, long process chains, and poor collaboration have always been the three major deadlocks in traditional site selection. Now, a single sentence can generate one's own dedicated site selection report. "At present, brands we work with, such as Baodao, Weila, and Xu Shengji, have become the first batch of LBSwork experience customers."

From a single location query to a business analysis report, and then to a complete map application, the value of LBSwork is amplified layer by layer: it can not only answer questions and provide reports, but also analyze enterprise needs and operating conditions, and give judgments and recommendations.

This ability to go deep into productivity scenarios and deliver implementable results is WorkBuddy's strength, and it is also the direction that LBSwork, which specializes in maps, is delving into.

In addition, Li Bin, general manager of Tencent Maps Location Services products, also announced the latest Gaia 2.0 on site.

As Tencent Maps Location Services' full AI matrix product and capability, Gaia 2.0 includes two product layers: Tencent Maps LBSwork and the spatiotemporal Harness foundation base, and ultimately provides efficient solutions to target customers across industries. "Last year, what we answered was how to build the platform. This year, what we need to answer is what customers can ultimately get," Li Bin said.

Behind this is insight into changes in user habits: to query a map, you no longer necessarily need to open a map app.

Previously, users were accustomed to finding locations within map apps. But now, more and more users are beginning to use map services within chatbots and agents.

In Zhong Xiangping's view, this trend makes map information acquisition more convenient and more ubiquitous; it is of great use to consumers, and at the same time places higher demands on industry practitioners.

At the same time, this also contains new opportunities for Tencent Maps: information acquisition entry points are broader; but actions based on maps, especially actions beyond lightweight needs, will still converge on professional platforms.

To capture new demand, Tencent Maps needs to advance more comprehensive and higher-level ecosystem openness.

Specifically, it hopes to prepare the capabilities, data, MCP, and Skills behind the map, so that agents in various scenarios can invoke them efficiently, flexibly, and at low cost, allowing users to consume fewer tokens and get better results.

On the other hand, various AI apps and agent platforms also need Tencent Maps' assistance.

The reason is that Tencent Maps holds the advantage of "maps."

Compared with general information queries, AI needs related to LBS often cannot be satisfied by data alone; users still very much need to "see clearly" where they are, where they are going, and what the route looks like. Graphical expression is crucial.

Zhang Zhidong gave an example: when Tencent Maps cooperated with OPPO, it found that in scenarios such as checking the weather and planning travel routes, AI answers with images had a click-through rate of over 30%; without images, it was only around ten percent.

In addition, after users obtain map information, subsequent actions still require the support of map services, and data needs to flow over a longer chain. This is also a link where Tencent Maps can leverage its strengths.

An open ecosystem with agentized maps as infrastructure is Tencent Maps' vision. As Zhong Xiangping said, in the AI era, Tencent Maps wants to become the most friendly map for the agent era and help all agents.

At the same time, Tencent Maps also has many areas that need further strengthening.

For example, it hopes to make the "readiness" of atomic capabilities higher, so that there can be a better experience in various scenarios.

Taking service completeness as an example, when cooperating with OPPO, Tencent Maps not only organized location information, but also provided integrated direct access to WeChat Mini Programs behind the locations and related surrounding services; taking response time as an example, in small and micro scenarios, users need more concise content, so it is usually necessary to ensure fast and accurate answers.

The same applies to the requirements of hardware manufacturers such as smart glasses—an extra 0.1 second of delay will make users feel it is "laggy."

Previously, Tencent Maps and Rokid launched the industry's first native ride-hailing service for AI glasses. Users do not need to take out their phones and can complete the entire process from calling a car and selecting a car to payment through conversation.

To make ride-hailing via AI glasses smooth, Tencent Maps turned capabilities such as payment, maps, travel, and content into standard components that can be directly invoked by the glasses. After the user issues a command, the AI glasses only need one invocation to activate multiple capabilities to execute the task.

This is exactly what Zhong Xiangping said: Tencent Maps does not make hardware, but hopes to become the "hands" and "feet" of smart hardware, so that partners can focus their resources and energy on innovation.

In the conversation, Tencent Maps management believed that it has currently achieved a passing score and is still developing toward excellence.

To achieve "excellence," the problem Tencent Maps is solving is to achieve the Pareto optimum among cost, token consumption, performance, and speed.

For example, in many scenarios, map services cannot simply pursue speed, but must also take detailed experience into account, so that users feel it is smoother and the information is more accurate.

Tencent Maps, which is becoming agentized, fully open, and continuously evolving, also has to face new commercialization paths and new value space.

"Special statement: The above content (including videos, pictures, or audio if any) is uploaded and published by users of "Dafeng Hao," a self-media platform under Phoenix Net. This platform only provides information storage space services.

Notice: The content above (including the videos, pictures and audios if any) is uploaded and posted by the user of Dafeng Hao, which is a social media platform and merely provides information storage space services."