WeChat is expanding its native AI assistant, "Xiaowei," with a new round of gray-scale testing that introduces AI entry points across three high-frequency usage scenarios. This update brings the total number of AI access points within the platform to 16, marking a significant step in embedding AI deeper into everyday interactions.
In the official accounts message feed, a new "AI Summary" icon with the Xiaowei logo has been added to the far left of the "Frequently Viewed Accounts" section. Tapping it triggers Xiaowei to compile recent updates from accounts the user frequently follows into a concise digest, streamlining information screening and delivery. Additionally, an "AI Summary" button has been added at the bottom of individual article detail pages, allowing users to instantly generate a summary of the current piece they are reading.
For the Moments feed, a long press on a friend's text post now reveals an "AI Comment" option. This feature enables Xiaowei to analyze the content and offer a range of commentary styles for users to choose from, providing a convenient starting point for engagement.
When viewing images within chat conversations, a new "AI Process" button appears at the bottom of the screen. This tool allows users to beautify images, extract text, or make modifications directly inside WeChat without needing to switch to a third-party app.
Beyond these three primary additions, WeChat is also quietly testing several other AI capabilities with a limited user base. For instance, in the Moments publishing interface, an "AI Write" button appears at the bottom right after uploading a photo. Once clicked, Xiaowei interprets the image content and generates multiple caption options at once. Users can select a preferred text, and it will be automatically filled into the posting field, eliminating the need for copy-pasting.
The "Scan" feature has also gained a new "Photo to AI" visual question-and-answer function. After capturing a photo, Xiaowei recognizes the content and suggests follow-up questions, such as identifying a brand. For packaging with foreign languages, options like "Translate," "Extract Text," and "Summarize Image Content" are displayed at the bottom, allowing users to complete recognition, translation, or further inquiries within a single session.
In the chat input field, a new "Command" feature has been introduced. Users can press and hold the input bar to speak, and a "Command" option will appear at the bottom. By sliding a finger to this area and stating a request, the AI interprets the intent upon release and directly generates a message suited to the current chat context, which can then be sent. This supports text polishing, translation, and message drafting.
At present, all these new functions are in the gray-scale testing phase, meaning only a randomly selected group of users can see them. Whether they will be fully rolled out to the general public is pending the official release from WeChat.