Summary of Key Points
Tencent is testing an AI Agent built into WeChat, which can be considered a “versatile assistant exclusive to WeChat.” The compliance approval process is expected to begin this month, and if approved, the service will first undergo a small-scale test before being gradually rolled out. This AI feature can be accessed by swiping right on the main WeChat interface and allows users to complete practical tasks using natural language, such as ordering coffee or calling a car. For example, if you request “a non-sweet coffee within 30 yuan that I can pick up myself,” the system will automatically search for relevant mini-programs, match them, and even place the order, integrating WeChat’s payment functions, mini-programs, and lifestyle services into a seamless experience. While WeChat has the natural advantage of having 1.4 billion monthly active users, it also faces competition from rivals like Alibaba and ByteDance. Before launching the service, they need to address challenges related to compliance, computing power, and ecosystem integration. In the future, there could even be new scenarios where user Agents interact with merchant Agents.
What is this AI Agent? — The “Task Assistant” within WeChat
Simply put, it’s not just a regular chatbot; it’s an AI that can actually help you complete tasks. For instance, if you want to order coffee, you don’t need to open the Starbucks or Luckin mini-programs one by one. You can simply state your requirements (price, flavor, pick-up option), and the AI will automatically search through all relevant WeChat mini-programs, filter the results, and even fill in the necessary information and complete the payment if authorized.
The access to this feature is very convenient—just swipe right on the main WeChat interface. This integration of various services (mini-programs, payment, lifestyle services) allows you to accomplish tasks without having to lift a finger.
WeChat’s Strengths for Developing an AI Agent — 1.4 Billion Users + Closed Ecosystem
WeChat is far more than just a chat tool; it’s also a payment platform, a content platform (with official accounts and video accounts), and a marketplace for mini-programs. Adding an AI Agent to these existing services activates all of its potential. With its massive user base of 1.4 billion monthly active users, WeChat can instantly reach everyone. Moreover, its ecosystem is already well-established, meaning other companies would need to seek permission from WeChat to integrate their AI services (for example, other apps would have to redirect users to WeChat mini-programs to use their AI ordering features).
Tencent has previously stated its goal of creating an AI that can connect mini-programs with lifestyle services, and this test is a step towards realizing that vision.
Why is Tencent Pushing This so Hardly? — Competition from Rivals
Rivals like Alibaba and ByteDance are already making significant progress in the AI Agent space. For example, Alibaba’s “Qianwen” can help with shopping, and ByteDance’s “DouBao” can handle various lifestyle needs, with their AI apps seeing rapid user growth. If Tencent delays, it might lose users to these competitors. Since WeChat is widely used on mobile devices, if other AI tools become more convenient, users could switch. Therefore, Tencent must accelerate the rollout of its AI Agent to retain its user base.
Three Major Challenges Before Launch
To get 1.4 billion users to use this AI, Tencent needs to overcome three major hurdles:
1. Compliance: Regulatory requirements state that generative AI services must be registered, and as a national-level app, WeChat faces stricter oversight. It must obtain the necessary permits before launching.
2. Computing Power: AI requires substantial computing resources to handle user requests efficiently (especially when 1.4 billion people are using it simultaneously). Tencent hasn’t accumulated enough NVIDIA chips like its competitors, and domestic chips are in short supply, posing challenges in terms of cost and stability.
3. Ecosystem Integration: WeChat has millions of mini-programs of varying quality. The AI Agent needs to ensure that merchant interfaces are reliable, services are trustworthy, and the payment process is smooth. It also needs to handle issues such as recommendation algorithms and profit distribution (for example, deciding which stores to recommend). A malfunction during the ordering process could ruin the user experience.
Possible Future Developments — Interaction between Agents
Once the AI Agent is launched, interactions might not be limited to users commanding the AI. Users may have their own Agents, and merchants might also have their own Agents (for example, a coffee shop’s AI could manage inventory and negotiate prices). The platform itself might also have its own AI to make decisions (such as determining which stores to recommend). For instance, when you request a coffee, your Agent would communicate with the coffee shop’s Agent to confirm available options, and the platform’s Agent would decide the order priority. However, this could lead to issues like Agents competing for orders or information chaos if not properly managed.
Final Note: Don’t Celebrate Just Yet
The WeChat AI Agent is still in the testing phase and must go through multiple stages, including approval, external testing, optimization of computing power, and ecosystem integration. While there’s a long way to go before widespread adoption, if successful, it could transform how we use WeChat by becoming a central hub for AI-powered tasks.