This week, WhatsApp, which is owned by Meta, updated its business API guidelines to prohibit general-purpose chatbots from operating on its platform. This decision is expected to impact WhatsApp-based AI assistants from companies such as OpenAI, Perplexity, Luzia (supported by Khosla Ventures), and Poke (backed by General Catalyst).
A new section has been introduced in the business API terms specifically for “AI providers,” with an emphasis on general-purpose chatbots. These updated terms, set to take effect on January 15, 2026, state that Meta will not permit providers of AI models to offer their AI assistants through WhatsApp.
Developers and providers of artificial intelligence or machine learning solutions—including, but not limited to, large language models, generative AI platforms, general-purpose AI assistants, or other similar technologies as defined solely by Meta (“AI Providers”)—are strictly forbidden from using or accessing the WhatsApp Business Solution, either directly or indirectly, for the purpose of providing, delivering, selling, offering, or otherwise making such technologies available when these technologies are the main (not incidental or secondary) feature being offered, as determined at Meta’s sole discretion.
Meta confirmed this policy change to TechCrunch, clarifying that it does not impact businesses that use AI to assist customers on WhatsApp. For example, a travel agency employing a bot for customer support will still be able to use the service.
Meta explained that the WhatsApp Business API was created for companies to interact with their customers, not as a platform for distributing chatbots. The company noted that while the API was intended for business-to-business scenarios, it has recently seen unexpected use cases involving general-purpose chatbots.
“The WhatsApp Business API was developed to help companies deliver customer service and send important updates. Our priority is to support the tens of thousands of businesses building these solutions on WhatsApp,” a Meta representative told TechCrunch.
According to Meta, the introduction of these chatbot use cases has significantly increased message traffic and placed new demands on their infrastructure, which the company was not prepared to handle. As a result, Meta is restricting use cases that do not align with the API’s original intent and strategic direction.
This change will essentially prevent WhatsApp from being used as a channel to distribute AI assistants or agents, leaving Meta AI as the sole assistant available on the platform.
Last year, OpenAI brought ChatGPT to WhatsApp, and earlier this year, Perplexity introduced its own bot to the app, aiming to reach WhatsApp’s user base of over 3 billion people. These bots could respond to questions, interpret media files, answer queries about them, reply to voice messages, and create images, likely contributing to a surge in message volume.
However, a larger issue for Meta was that the WhatsApp Business API is a key revenue source, charging companies for various message types such as marketing, utility, authentication, and support. Since the API was not designed to accommodate chatbots, WhatsApp could not monetize these interactions.
During Meta’s Q1 2025 earnings call, Mark Zuckerberg highlighted that business messaging represents a significant revenue opportunity for the company.
“Currently, most of our revenue comes from ads on Facebook and Instagram feeds,” he stated. “But WhatsApp now has over 3 billion monthly active users, including more than 100 million in the US, and that number is rising quickly. Messenger is also used by over a billion people each month, and the daily message volume on Instagram now matches that of Messenger. We believe business messaging will become a major part of our business.”
