Meta has decided to restrict the use of WhatsApp’s business API by independent AI chatbots. The change, effective January 15, 2026, will prevent companies like OpenAI, Perplexity, Luzia, and Poke from offering assistants directly on their chat apps.
The new rules will add a section for AI developers to the WhatsApp Business Policy. It defines them as providers of large-scale language models or generation systems and prohibits the distribution of chatbots that function as general assistants. The company says the API is intended for business communication and not for deploying public AI tools.
WhatsApp refocuses on business use
Meta designed the Business API to help businesses manage customer messages, confirmations, and product updates. These features are not affected. What is changing is the growing presence of chatbots that perform a wide range of conversational tasks. OpenAI and Perplexity both launched Assistant on WhatsApp earlier this year. They can answer questions, read images, and reply to voice notes, which increased traffic. Mehta said the system wasn’t built to handle that kind of workload or the support demands that come with it.
Earnings structure in crisis
Beyond the strain on the system, there are economic reasons for the change. The Business API generates revenue for Meta by charging businesses based on message categories such as marketing and support. The chatbot didn’t fit that model, so Meta couldn’t collect fees from its activity. By removing them, the company protects one of the few direct revenue channels within WhatsApp.
Mark Zuckerberg said business messaging is a key area for future revenue. WhatsApp already has more than 3 billion users worldwide and continues to grow in the United States. By limiting AI chatbots, Meta keeps the platform focused on its commercial core and away from external experimentation that could obscure its purpose.
Policy language adds new restrictions
The updated terms go beyond blocking chatbots. Companies cannot use WhatsApp data to train or improve AI models, even in anonymous form. Although you can fine-tune internal models for personal use, Meta prohibits sharing that data or using it to develop large-scale systems. The Company reserves full authority to determine what qualifies as an AI Provider and may terminate any account that violates these Terms.
The larger strategy of the meta
This move ensures that Meta AI remains the only assistant built directly into WhatsApp. This also signals increased control over how other developers can interact with the platform. For AI companies, losing access to WhatsApp means losing a valuable channel for billions of users. For Meta, this is a step toward keeping innovation and revenue within its own network rather than having external models grow on its own infrastructure.
Note: This post was edited/created using GenAI tools. Image: DIW-Aigen.
Read next: Wikipedia faces decline in human traffic as AI and social video change search habits
