Meta has released Muse, a personal AI agent that helps users shop, book services, manage calendars, and complete other tasks (Retailgentic). Users can access Muse at muse.ai, log in with Facebook or Instacart credentials, and customize a personal agent avatar. The agent operates as a browser-based tool that navigates websites, filters products based on user criteria, and presents curated product cards for selection.
Muse offers two purchase paths: users can authorize the agent to complete transactions directly using a Stripe Link token for secure guest checkout, or they can take over the browser themselves to finish the purchase on the merchant website (Retailgentic). For commerce practitioners, this launch is significant because Meta brings distribution scale—the company operates four products with over 1 billion users each—giving Muse a consumer adoption advantage that earlier shopping agents like OpenClaw, Hermes, and GrokBot lacked. Merchants must now prepare for AI agents navigating their sites, ensure website compatibility with automated browsing, and consider how to optimize checkout flows for agent-initiated transactions.
The agent's browser-based approach reveals both capabilities and limitations: it successfully navigates multiple retailer sites and filters by price and product specifications, but struggles with poorly structured websites and missing technical specifications. This suggests merchants with clean, accessible site architecture and complete product data will perform better with AI agents, making web optimization a new competitive priority in the agent-commerce era (Retailgentic).