Muse can shop, write emails, and negotiate prices for users, all through WhatsApp

What it does

Meta has launched Muse, an AI agent that operates entirely through WhatsApp to perform a range of practical tasks. Muse can book travel, handle online purchases, negotiate prices, and write emails. It supports transactions by integrating payment through Stripe’s Link, allowing users to pay directly within WhatsApp. Additionally, Muse is paired with a security agent called Sentinel, which reviews every action before it hits the internet to prevent risks.

Why it matters

Muse upgrades AI agent functionality by embedding commerce, communication, and negotiation into the most widely used messaging platform globally. This lowers the barrier for users and businesses to automate complex workflows without leaving WhatsApp. Unlike OpenAI’s ChatGPT, which has discontinued its direct checkout feature, Meta embeds payment processing natively. That pushes a tighter link between AI-driven shopping and real-world transactions, accelerating digital commerce through conversational AI. Adding a monitoring layer in Sentinel addresses increasing concerns around AI mistakes, misuse, or security vulnerabilities by vetting requests pre-execution.

Who it is for

Muse targets everyday users who want to automate routine tasks in the channels they already use, like booking flights or negotiating prices for goods/services through chat. For businesses, this expands commerce availability on WhatsApp, which hosts over 2 billion users, potentially increasing sales and reducing friction. Developers and AI service providers will watch Muse as an example of integrating AI-driven agents with payments and real-time negotiation, pushing operator expectations around multi-functional conversational agents.

The catch

Muse’s capabilities hinge heavily on Meta’s control of WhatsApp and Stripe’s payment infrastructure, which may limit replicability in other ecosystems. Dependence on Sentinel as a security monitor raises questions about how robust the filtering is against errors or malicious commands. The negotiation feature may still face challenges in handling complex buyer-seller dynamics reliably and avoiding user frustration. While integration is seamless, adoption depends on how comfortably users trust an AI doing sensitive tasks like payment and negotiation without human oversight.

What to watch next

Monitor how Muse’s security system Sentinel scales as usage grows and whether it effectively prevents fraud or mishaps. Watch if similar AI commerce bots expand beyond Meta’s ecosystem, and how OpenAI and others respond to the competitive pressure from embedded payments. This could shift how conversational AI tools prioritize real transaction functionality over chat assistance. Tracking Muse’s adoption rate on WhatsApp will reveal how willing users are to delegate negotiations and purchases to AI agents in everyday messaging.

AI Quick Briefs Editorial Desk