Tencent's WeChat team is secretly developing a high-priority AI Agent (intelligent agent) product, aiming to be fully integrated into the millions of mini-programs within the WeChat ecosystem. The project is led by WeChat's technical director Zhou Hao, who reports directly to Zhang Xiaolong, and is scheduled to start in the first half of 2025.
According to the information, the Agent is planned to begin a phased testing in mid-2026 and gradually expand in the third quarter. In terms of core functions, the product aims to achieve cross-app task execution capabilities, allowing users to call third-party services such as ride-hailing, food delivery, ticket booking, and payment through natural language commands, upgrading WeChat from a "social entry point" to an "intelligent task center."
In terms of technology selection, the WeChat team has maintained a cautious and open attitude. Although Tencent has recruited researcher Yao Shunyu from OpenAI to lead the upgrade of the Hunyuan large model system, the WeChat Agent has not fully bet on Hunyuan so far. Instead, it is simultaneously testing external large models such as Zhipu, Alibaba, DeepSeek, as well as its own small-scale self-developed models, to ensure stability in handling complex multi-step tasks.
This move is seen as a key counterattack by Tencent in the field of AI applications. Previously, although Tencent launched an independent app "Yuanbao," it still lags behind ByteDance's "Doubaobao" and Alibaba's "Tongyi" in terms of monthly active users.
The deep involvement of WeChat in the Agent market marks that the competition of large models has shifted from "dialogue and question answering" to "execution and collaboration." With the advantage of 1.4 billion monthly active users and a steadily maturing mini-program ecosystem, WeChat Agent has the potential to reshape the operational logic of mobile internet. If it successfully overcomes the technical barriers of integrating external models and accessing private data, WeChat is highly likely to become the most widely covered AI intelligent assistant platform in China, thus controlling the traffic entry point for the next generation of AI-native applications.

