The trend of "AI assistants" has finally reached WeChat. On March 22, WeChat officially launched a new plugin called ClawBot, enabling users to connect to and control OpenClaw directly through the chat interface. This allows for remote device operation, file handling, and schedule management. While not immediately available to all users, the feature is being rolled out gradually. However, it appears WeChat was not fully prepared for this release. Many users note that "WeChat ClawBot" falls short of expectations. For instance, it cannot be added to group chats, interactions remain largely text-based, and handling complex tasks is still challenging.

WeChat is famously restrained. Any new feature undergoes extensive deliberation and internal testing before launch. Over the years, its core interface has remained largely unchanged. Last year, it even restricted Doubao AI Assistant and imposed limitations on its own Yuanbao AI. Faced with the "AI assistant" trend, WeChat seems to have finally broken its usual caution. But why the urgency this time? And what does this mean for us?

The reason may lie in challenges to WeChat's competitive moat. What is that moat? It's WeChat's over 1.3 billion users and the powerful network effects they create. A single WeChat account holds little value; ten accounts offer minimal worth. But ten million accounts are a different story. Each new user adds value for all existing users. With 1.3 billion users, WeChat's value becomes immense due to network effects, making it a foundational connectivity platform for the era.

This deep moat allows WeChat to accomplish feats others cannot. For example, it initially avoided mobile payments, but the viral success of "WeChat Red Packets" during a single Spring Festival reportedly achieved the same bank card linkage volume that took Alipay a decade. Similarly, many Tencent games succeed largely due to WeChat's massive distribution and social sharing network. Thus, WeChat has maintained its own pace amid various industry trends—whether short video wars, food delivery battles, or the metaverse—as long as its moat remains unchallenged.

So why is the moat now under threat? Because of "AI assistants." Recently, OpenClaw gained significant popularity. Behind this trend lie four distinct layers: 1) The Model, which determines the AI's intelligence and emotional value; 2) Skills, enabling specific tasks like songwriting or flight booking; 3) The AgentOS, the core system that interprets complex commands and coordinates skills; and 4) The Interface, the platform for user interaction, such as Telegram or WhatsApp.

WeChat can afford to ignore the first three layers, but the Interface is critical—it is part of its moat. In recent months, many users, including long-time contacts, have reactivated accounts on Telegram and WhatsApp primarily to use AI assistants. This means valuable, active users are spending more time on alternative platforms, which also offer human communication. Some platforms, like Telegram, are particularly well-suited for AI interactions. Hence, WeChat felt compelled to act.

However, WeChat's cautious approach is evident in its implementation. "WeChat ClawBot" operates like a special zone, isolated from the main ecosystem. Articles, images, and chat histories cannot be forwarded to it, maintaining a strict separation from core WeChat functions. It seems the AI trend has unsettled not just users but platforms as well.

What implications does this hold for us? The key takeaway may be the four-layer model: Model, Skills, AgentOS, and Interface. Understanding this framework helps decipher developments around AI assistants. For instance, WeChat's current focus is primarily on the Interface—securing the user gateway—rather than developing superior models, skills, or operating systems. Thus, significant functional improvements should not be expected near-term.

Where do opportunities lie for individuals and smaller players? Competing in interfaces is difficult against WeChat; developing large models is a game for giants; and creating an AgentOS requires deep technical expertise. The real opportunity lies in Skills. If you excel at creating presentations, consider developing a skill that generates PPTs from a single command. Legal experts can build specialized consultation tools, while travel enthusiasts might design skills for customized itineraries. A vast market for Skills could emerge, akin to today's App Store, offering niches for individuals and small teams.

Admittedly, the pace of AI innovation is dizzying, and feeling overwhelmed is natural. Yet, the antidote to anxiety is understanding. Recognizing the underlying patterns in complex changes reduces fear and reveals hidden opportunities. This may be the essential lesson from every major shift: comprehend first, then commit fully.