Meta is accelerating development of its personal AI agent, Muse, introduced just weeks ago. CEO Mark Zuckerberg emphasized Muse as a central part of Meta’s future during the company’s annual Connect event on September 23, 2026. Muse aims to assist with everyday tasks by integrating deeply with users’ apps and digital services like email, calendars, and soon, even Meta’s AI glasses. Powered by Meta’s own multimodal AI model called Muse Spark, the agent connects seamlessly across devices to handle busy work and improve productivity.

One of the most notable new features previewed at Connect is the introduction of personalized digital avatars for Muse, enabling users to video chat and interact in real-time with their AI. These avatars, powered by the new Muse Realtime Avatar model, bring a face, voice, and personality to the previously faceless assistant. This adds a more human touch to AI conversations and task management, reinforcing Meta’s vision of making Muse a superintelligence that billions will rely on to accomplish personal and professional goals.

Meta also announced that Muse will soon integrate with its smart glasses line, allowing users to engage with their AI assistant by voice commands directly through wearable devices. Muse will assist with activities ranging from workouts to booking appointments and shopping, operating discretely in the background and updating users upon task completion. The company expects this feature to be ready in the forthcoming months, expanding Muse’s presence beyond phones and computers.

Further expanding Muse’s capabilities, Meta revealed plans to give Muse its own email address to autonomously handle messages and manage threads. The AI agent will also be able to control Mac devices to perform work tasks independently. Additionally, Meta has partnered with major retailers and payment platforms such as Stripe, Shopify, PayPal, Best Buy, Gap, Sephora, Walmart, and Wayfair to enable shopping and transactions through Muse. The company has opened its platform for third-party developers to build connectors, receiving strong interest with over 1,500 applications, aiming to create a broad ecosystem around their AI assistant.