Meta built a wall around WhatsApp and kept every rival AI assistant outside. The European Commission knocked that wall down. In October 2025, Meta announced it would ban all third-party AI chatbots from WhatsApp starting January 15, 2026. Only Meta’s own assistant would remain accessible. The company barred rivals from using the WhatsApp Business API, the tool that lets outside services connect to the platform.
The European Commission opened an abuse of dominance investigation. After Meta refused to budge, the Commission issued interim measures on June 9, 2026. The order required Meta to restore free access to rival AI assistants within five working days.
EU antitrust chief Teresa Ribera did not mince words. She said Meta expected to “leverage the vast reach and likely dominance of WhatsApp to benefit its own AI assistant and to foreclose rivals.”
Meta called the order regulatory overreach and promised to appeal. The company argued that the EU was letting OpenAI and other large firms use a paid product for free.
Meta faces fines of up to 10% of its global annual turnover if it breaches EU antitrust rules. The interim order stays in place until the investigation ends or until June 2029, whichever comes first.
Users Can Now Add Five AI Assistants Inside WhatsApp
WhatsApp began rolling out third-party AI agent support on Android in September 2026. The feature lets users connect up to five external AI services directly inside the app.
These agents come from independent developers. They include tools from OpenAI, Anthropic, Google, and smaller companies.
The setup works through a simple system. You add an agent through WhatsApp settings and give it a name and profile photo. WhatsApp generates an API key that links the agent to a dedicated chat. You then talk to that agent like any other contact.
Each agent can only see messages in its own conversation. It cannot read your other chats, contacts, or media files.
WhatsApp does not allow these agents in group chats. That restriction prevents exposure to sensitive multi-party conversations.
A Privacy Tradeoff Every User Should Understand
WhatsApp built its reputation on end-to-end encryption. Messages between two people stay private. Not even WhatsApp can read them.
Conversations with third-party AI agents do not carry that protection. Messages sent to these agents are not end-to-end encrypted because they route through a Meta-managed service on behalf of the agent developer.
That means the agent developer can access messages in unencrypted form. Meta’s own Business Agent carries the same caveat. Alice Newton-Rex, head of product at WhatsApp, confirmed that interactions with WhatsApp AI agents are not end-to-end encrypted and that message data could potentially be used for AI training.
Regular WhatsApp chats remain encrypted. The change only applies to conversations with AI agents.
Users should review each provider’s data policies before sharing sensitive information. This privacy gap represents a real cost of convenience.
Meta Still Plans to Make Money From AI on WhatsApp
Opening WhatsApp to rivals does not mean Meta gave up on AI revenue. The company has a clear strategy to monetise the platform.
Meta launched its Business Agent globally in June 2026. The tool handles customer conversations on WhatsApp, Messenger, and Instagram. It answers questions, recommends products, books appointments, and closes sales without human involvement.
Businesses pay for this service. Meta charges $2.00 per million tokens for its native agent, which works out to roughly four to five cents per message. That pricing took effect on August 1, 2026.
Meta also charges for service messages sent through the WhatsApp Business Platform. Starting October 1, 2026, businesses pay per message for customer service replies inside the 24-hour window. Those replies used to be free.
The revenue potential is enormous. WhatsApp’s business messaging revenue hit a $2 billion annualized run rate in late 2025. That sounds large until you compare it to Meta’s total sales of $200 billion that year.
Meta sees AI agents as the key to closing that gap.
Why This is Beyond WhatsApp
The EU’s intervention sets a precedent for how regulators treat AI access on dominant platforms.
The Commission used interim measures for the first time since 2019. Ribera made clear the EU would act again if similar circumstances arise.
This case matters for three reasons.
First, it establishes that dominant messaging platforms cannot pick winners in AI. If a company controls the pipes, it must let rivals compete on equal terms.
Second, it protects smaller AI developers. Startups like Poke and Luzia filed complaints with EU regulators when Meta blocked them. The interim order gives them access to WhatsApp’s user base without needing special deals.
Third, it shows regulators will move fast when AI markets are at stake. The Commission acted because AI markets develop exceptionally fast. Waiting for a full investigation could have let Meta lock in its advantage permanently.
Meta may appeal, but the direction is set. WhatsApp is becoming a platform where AI assistants compete on quality rather than access.
What Comes Next
Meta’s decision to open WhatsApp to third-party AI agents looks like a loss of control for the company. In reality, it strengthens the platform’s long-term position.
Users get more choices without leaving their favourite app. Developers get access to a market they could only dream about before. Meta gets a clearer path to monetisation that does not depend on locking out rivals.
The era of the walled garden is ending. WhatsApp’s users will be better off for it.
