WeChat, the messaging service used by more than a billion people, is opening its most guarded asset: the ability to place calls and send messages. The company said this week it is working with handset makers Huawei, Honor and Xiaomi to let their built-in AI assistants initiate WeChat voice and video calls and send messages directly, without the user opening the app.
The feature, which the companies describe as an A2A capability, lets a system-level assistant act on a spoken request, find a contact and place the call through WeChat’s network. A user can ask the phone’s assistant to call a friend or send a message, and the assistant does it end to end, reading back the action for confirmation before anything is sent.
The significance is the boundary being crossed. WeChat has historically guarded its communication functions, insisting that calls and messages happen inside its own app, where it controls the experience and the data. This is the first time the company has exposed its core communication capabilities to external AI entry points, a structural change in how one of the world’s largest social platforms interacts with the rest of the phone.
The motivation is defensive. Phone makers have been building AI assistants into their devices with growing ambition, and the fear inside WeChat, according to people familiar with its thinking, was that assistants would one day route around the app entirely, letting users reach people through other channels. By partnering with the handset makers, WeChat keeps itself at the center of the transaction, even as the interface shifts from the app to the assistant.
The deal also answers a competitive question. The phone AI assistants, powered by models from domestic and international labs, are becoming the default way many users interact with their devices. A system assistant that cannot touch WeChat, the most used app in the country, is a poor assistant; one that can place WeChat calls becomes essential. By opening up, WeChat ensures the assistants need it rather than replace it.
The technical details matter. The A2A capability works through a protocol that lets the assistant request actions from WeChat’s servers, with confirmation steps built in to prevent accidental sends. The companies said privacy protections apply, with the assistant given access only to the actions the user authorizes. WeChat’s position as a regulated communications service means the feature must also satisfy the same oversight that applies to its own calls.
The partnership list is notable. Huawei, Honor and Xiaomi account for a large share of smartphone sales in China, and each runs its own assistant platform. By working with all three rather than one, WeChat avoids dependence on any single handset maker, keeping the balance of power with the platform. Other manufacturers are expected to be added later, according to people familiar with the plans.
The move mirrors a global shift. In the West, Apple and Google are embedding AI assistants into their operating systems, and app makers are being asked to expose their functions to those assistants through similar protocols. WeChat’s decision to participate is a recognition that the phone’s operating system, not the app, is becoming the center of the digital experience, and that platforms which refuse to be reached by assistants risk being left out of the conversation entirely.
The commercial stakes are clear. WeChat generates revenue through payments, mini-programs and advertising, all of which depend on users spending time in the app. An assistant that places calls through WeChat, and asks before sending anything, keeps the app central to daily communication while shifting the entry point. The company is betting that opening its walls keeps its moat intact.
WeChat’s scale makes the decision consequential. The service has more than a billion users and is the default communication channel for much of daily life in China, handling payments, news, appointments and commerce alongside messaging. Any change to how its core functions are reached ripples through the country’s digital economy, which is why the company has been conservative about exposing its functions to outside systems. The A2A opening is a carefully controlled exception, and the list of approved partners is part of the control.
The technical approach is worth noting. Instead of giving each handset maker a separate integration, WeChat has defined a single interface that any approved assistant can use, a design that keeps the company in the middle of every transaction. The same pattern is emerging globally, with phone makers and app developers standardizing how assistants invoke app functions, and WeChat’s move puts it in that standard’s foundation in its home market. The company that defines the interface defines the ecosystem.
Analysts see the move as both defense and offense. Defense, because a phone assistant that could not reach WeChat would push users toward rival apps that could; offense, because WeChat’s position as the approved destination makes it the default channel for assistant-driven communication, protecting its role as the entry point for commerce and social interaction. The company has made similar calculations before, opening mini-programs to developers only after the risks were contained. The A2A launch follows the same playbook, and the next test will be how far the opening extends.
For users, the feature is a convenience with a learning curve: trusting an assistant to place calls and send messages requires confidence that it understands context and will not misfire. The confirmation step addresses that, and the companies said the feature would roll out gradually. If it works, the quiet revolution is this: the most important app on the phone is starting to answer to the phone itself.
