en.Wedoany.com Reported - On April 23, Tencent held the 2026 TIME DAY·Tencent Smart Mobility Technology Open Day in Beijing, officially unveiling the "Mobility Full-Scenario Agent Open Platform" for the automotive industry and upgrading its cabin agent solution. Tang Daosheng, Senior Executive Vice President of Tencent Group and CEO of Cloud & Smart Industries Group (CSIG), stated during the event that the automotive industry competition is being redefined by AI. The gap between automakers in the future will no longer hinge on whether they possess intelligence capabilities, but on who can implement AI at scale and systematically. Automotive intelligence is not just about "putting AI in the car," but "reconstructing the enterprise with AI"—from R&D collaboration and office workflows to marketing conversion and customer service, traditional human-dependent processes are being systematically reshaped by AI.
The Mobility Full-Scenario Agent Open Platform adopts a three-tier architecture: the foundational capability layer, the platform capability layer, and the application ecosystem layer. The foundational capability layer fully upgrades the end-cloud collaborative cabin large model architecture, deeply integrating core capabilities such as Yuanbao Search, WeChat Pay, and spatial intelligence. The platform capability layer provides a full-stack toolbox for agent orchestration, extension, and operation, allowing automakers to flexibly invoke capability components from different layers, as if selecting products as needed from a shelf. The application ecosystem layer targets high-frequency user scenarios such as commuting, travel, entertainment, and charging, launching seven core scenario agents: On-the-Go Ordering, On-the-Go Guide, On-the-Go Strolling, On-the-Go Connectivity, Owner Services, Video Entertainment, and News & Information, aiming for out-of-the-box cabin agent deployment.
Among the seven agents, On-the-Go Ordering covers the entire service process from selection and ordering to meal pickup. It can filter the best store based on the user's real-time location and preferences, and inversely calculate the meal preparation time based on the vehicle's estimated arrival time, ensuring the meal is ready just as the user arrives. In long-distance travel scenarios, the On-the-Go Guide acts as an AI tour guide, providing real-time push notifications of scenic spots and travel tips along the route. It also introduces a new fleet coordination feature supporting location sharing, real-time voice chat, and service linkage for multi-vehicle convoys. On-the-Go Connectivity integrates the OpenClaw capability. When a user sends a combined command in a WeChat dialog, such as "There's a passenger in the front seat. When the car starts, turn on the seat massage, play a welcome message, and set WeChat in the car to privacy mode," the in-car CarBot can automatically decompose the task and execute each step.
During the open day, Zhong Xuedan, Vice President of Tencent Smart Mobility, pointed out that the second half of automotive intelligence competition is not about having more functions, but about who can organize large models, vehicle capabilities, and service ecosystems into a perceptible, executable, and evolving intelligent hub. In response to the question of OpenClaw being directly integrated into the vehicle, Li Bo, also Vice President of Tencent Smart Mobility, stated that there are currently no plans to put it directly in the car. The core issue lies in access control. Until safety and permission issues are properly resolved, directly transferring external agents into the vehicle is inappropriate. A more reliable path is to issue commands to the in-car system through phones or cloud platforms after security processing.
Data on market coverage and automaker partnerships were also disclosed during the open day. Tencent has served over 100 automakers and mobility technology companies, with its intelligent cabin solutions penetrating more than 80% of leading auto manufacturers, cumulatively equipped in more than 18 million vehicles. Its smart driving cloud map scale has achieved three-fold year-on-year growth. In terms of deep AI collaboration, Tencent has engaged in joint exploration with over 40 automakers, including Changan, SAIC, and GAC, to promote the application of agent technology in specific scenarios. Regarding overseas deployment, Tencent provides overseas cloud services to over 50 major Chinese automakers and autonomous driving technology companies, covering local cloud deployment and compliance support.
At the technical foundation level, Tencent has built a full-chain capability from model development and engineering to application deployment. The latest Tencent Hunyuan model, Hy3 preview, achieves significantly improved complex reasoning and agent capabilities with lower activation parameters, as validated in products such as Yuanbao, WorkBuddy, and ima. Tencent Cloud provides full-chain intelligent computing solutions for automakers, achieving a mixed computing cluster utilization rate of 98.4% and improving GPU efficiency by over 60% in inference scenarios. For internal automaker operations, Tencent Cloud's CodeBuddy has been applied to requirements analysis, code generation, and test review for several automakers. FAW-Volkswagen, after integrating the large model, saw a reduction of over 25% in sales lead costs.
Regarding privacy and security design, the use of data from sensors like in-car cameras is conditional upon desensitization and declassification. Agents operate in sandboxed environments, with data operations strictly limited to a reasonable scope. Tencent emphasized that the secure operation of agents requires a full-chain protective system covering sensor data collection, transmission encryption, compute isolation, and access control, which is crucial for winning user trust and facilitating the industrialization of automotive intelligence.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
