Meta tests AI agent integration on WhatsApp
Meta is developing features to allow users to communicate with AI agents within WhatsApp. However, these conversations will not be end-to-end encrypted.

Meta is actively testing features that will enable users to interact with AI agents directly on WhatsApp. According to reports from WABetaInfo, which tracks WhatsApp's beta versions, the platform is exploring ways for users to connect with artificial intelligence entities within the messaging app.
The new functionality reportedly allows users to link up to five AI agents for personal use via API keys. A key detail in this rollout is that chats with these AI agents will not feature end-to-end encryption (E2EE). The agents are confined to individual chats and cannot be added to group conversations, thus preventing exposure of sensitive, multi-party discussions. Each agent maintains isolated access to its chat and cannot access other conversations, contacts, or media.
This move positions WhatsApp alongside competitors like Telegram, Discord, and Slack, which already offer similar AI integration capabilities. It aligns with Meta CEO Mark Zuckerberg's broader ambition to evolve WhatsApp into a "super app," akin to China's WeChat, by incorporating a wider range of services. This development also aims to keep WhatsApp competitive as other messaging platforms, such as Apple's iMessage and Google's messaging app on Android, increasingly embed AI tools.
The feature is currently being rolled out to a limited number of users, and it is unclear when Meta plans to make it universally available. Previously, developers relied on unofficial workarounds, such as reverse-engineered bridges, to connect AI agents to WhatsApp, methods that carried risks of account restrictions and security vulnerabilities.