Meta is loosening its grip on AI inside WhatsApp

Under the new arrangement, AI developers will be able to offer general-purpose chatbots — including services similar to ChatGPT or Claude — directly inside WhatsApp. But there’s a major condition: they’ll need to pay a usage fee for each message sent.

The cost ranges between €0.0490 and €0.1323 per non-template message, depending on the country. Because conversations with AI assistants typically involve dozens of back-and-forth messages, the pricing structure could quickly become expensive for developers trying to scale chatbot services.

This temporary policy shift is essentially Meta buying time. The company says allowing AI chatbots for a year should give regulators enough space to complete their broader antitrust investigation.

Interestingly, the rule doesn’t affect businesses using AI for customer support. Companies can still run automated service bots that send template responses through the WhatsApp Business API. The restriction — and the new paid access — primarily targets general-purpose AI assistants competing with Meta’s own AI offering.

That competitive angle is exactly why regulators stepped in. Governments in Europe, Italy, and Brazil opened investigations after Meta introduced the original policy last October, arguing it could unfairly block rivals while Meta promotes its own assistant, Meta AI, inside WhatsApp.

Meta has previously defended the restriction by saying that large AI chatbots place heavy demand on its systems — something the Business API wasn’t originally designed to handle.

But for the next year at least, Europe’s AI companies will get their chance to build AI-powered conversations inside the world’s biggest messaging app — as long as they’re willing to pay for every reply.