# Meta Launches Muse, AI Agent That Can Send Emails, Book Tickets and Shop With Your Approval > Source: > Published: 2026-09-08 22:11:58+00:00 # Meta Launches Muse, AI Agent That Can Send Emails, Book Tickets and Shop With Your Approval ## Discover how Meta's Muse AI agent is transforming personal task management with advanced privacy features Meta has launched Muse, a personal AI agent that can send emails, book travel, shop online and carry out other tasks on a user's behalf. The company says Muse can continue working in the background and return when it needs approval for sensitive actions, such as sending an email or making a purchase. The launch marks a shift in how Meta wants people to use AI. Rather than requiring users to guide a chatbot through every step, Muse can open a browser, fill in forms, organise plans and negotiate on a user's behalf. It is initially available in the US to people aged 18 and over through a dedicated app, the web and WhatsApp, with support for [Meta's AI glasses](https://www.ibtimes.co.uk/meta-ai-glasses-cinema-crackdown-could-penalise-blind-filmgoers-alongside-pirates-1815503) planned. But the biggest question is whether people will trust Meta with enough personal access to let Muse act on their behalf. The agent can connect to email, calendars, shopping, payments and other services, putting [privacy and security at the centre of the launch](https://www.ibtimes.co.uk/meta-ai-model-hacks-systems-security-test-1812825). According to Meta, Muse was built around those concerns, with a dedicated virtual machine and separate security system designed to keep its actions under control. ## What Meta's Muse AI Agent Can Actually Do Muse is designed to handle tasks that traditionally required users to move between apps and websites. Users can ask it to send emails, book travel, complete forms or search for and purchase products. For longer tasks, Muse can continue working after the user closes the app. Meta says it can create a personalised plan, coordinate resources and continue until it needs the user's input. Shopping is a key capability. Muse can complete purchases using Link by Stripe, which generates a single-use card number without exposing the user's normal payment-card details to the merchant. Meta says users must approve purchases, while Shop Pay and 1Password integrations are planned. The approval requirement is important because Muse can take actions with real-world consequences, rather than simply providing recommendations. ## Meta Builds a Security Layer Around Muse Meta says Muse runs inside a dedicated cloud-based computer called Muse Secure VM, which keeps the agent and users' connected data separate from other agents. A second system, Sentinel, monitors Muse and controls its internet access. Meta says it can block potentially dangerous activity and request approval for sensitive actions. Muse also cannot directly access passwords or payment details stored in secure systems. Muse provides an audit trail showing what it has done and intends to do. Users can choose connected services, set access levels and disconnect them later. Meta plans to introduce Muse Confidential VM later this year, encrypting the entire virtual machine, including conversations and personal data, with a key held only by the user. Meta says it has worked with external security researchers through a bug-bounty programme, while Confidential VM is being made available to external auditors. ## Why Muse Is Meta's Bid To Dominate the AI Agent Market The launch is part of [Mark Zuckerberg's broader push towards what he calls personal superintelligence](https://www.ibtimes.co.uk/mark-zuckerberg-meta-personal-superintelligence-1813323). In his vision, AI agents would operate continuously across areas such as work, finances, health, relationships and home management. That makes Muse strategically different from a conventional chatbot. Meta is trying to move AI from a system that responds to commands to one that can complete tasks on a user's behalf. The company has already been moving in this direction, with Meta saying its Muse Spark model could make plans, connect to email and calendar applications and handle tasks on a user's behalf. ## Can Meta Overcome Its Privacy Reputation and Win User Trust? The technology also creates a difficult trust problem for Meta. Giving an AI access to email, shopping and payments means giving it far greater control over personal information than asking a chatbot a question. Meta is therefore making privacy and user approval central to the product's pitch. The company says Muse does not share users' conversations or data stored in its secure environment directly with its advertising systems, although activity carried out by Muse on the web could indirectly influence the ads users see. Users can also tell it to forget information it has learned. Given the scrutiny Meta has faced over privacy and data protection, the broad access required for Muse is likely to attract close attention from regulators and advocacy groups. That attention could shape how quickly the agent expands beyond its initial US rollout and how freely users are allowed to connect sensitive accounts. Muse is initially rolling out in the US on iOS, Android and the web, with WhatsApp access also available. Meta says it will be free for most users, with subscription plans for people who want to use it more heavily. © Copyright IBTimes 2026. All rights reserved.