Exhibitor login
AI Insider 10 September 2026

Meta launches AI agent for groceries and payments

Meta launches AI agent for groceries and payments

Meta has launched a new AI agent that enables users to order groceries, compose emails, and make payments. This assistant, based on the company's Llama models, is currently available in the United States with plans for future expansion into other markets. This innovative system operates through Meta's platforms such as WhatsApp, Instagram, and Messenger. However, the introduction of the agent highlights the pressure that this technology places on existing privacy regulations, including the General Data Protection Regulation (GDPR) and the European AI regulation.

The AI agent is designed to execute tasks independently, meaning that users do not need to enter each command separately. Meta aims to facilitate the use of its platforms by combining multiple actions within the same interface. With clear consent and access to the necessary accounts, users can streamline tasks such as turning a grocery list into an order or processing payments. However, this makes it essential for Meta to adhere to strict privacy and security measures to ensure data protection.

The rollout of the AI agent in the US begins with a limited number of features and partners. The European launch is uncertain due to complex privacy laws and payment regulations on the continent. The GDPR requires companies like Meta to provide transparency about how data is collected and stored. This also applies to EU legislation regarding AI, which imposes additional obligations on organizations offering generative AI systems.

The implications for Dutch users are currently minimal, as the introduction of this technology starts outside the EU. Online retailers and other services have the opportunity to collaborate with Meta to integrate the agent but must comply with the GDPR and existing payment legislation. Governments and public institutions should exercise caution when using AI agents for critical tasks, considering the potential impact and the need for thorough evaluations and limits on transparency and control.

Read the full article from AI Insider.