Meta launched Muse on September 8, 2026, its first consumer AI agent product, rolling out to US users across iOS, Android, and the web at muse.ai. The company had been building toward this moment for over a year, and the arrival marks a genuine milestone: an AI that doesn’t just answer questions but actually carries out tasks on your behalf, at consumer scale.

This is the product that was previously referred to internally and by press as “Hatch.” It shipped under the name Muse.

What Muse Actually Does

Muse is designed to act, not just respond. It can book movie tickets, schedule appointments, handle online shopping, fill out permission slips, and work through multi-step tasks that used to require you to juggle several apps at once.

It runs inside a dedicated Muse Secure VM, essentially a sandboxed environment with its own browser, so the agent can access websites and services on your behalf without touching your personal accounts directly. Users grant specific permissions, and Muse operates within those bounds.

The interface is conversational and long-running. Unlike a standard chatbot where each message starts fresh, Muse remembers your context over time, learns from ongoing conversations, and gets more useful the more you use it. You can name your agent, set an avatar, and choose how it communicates with you.

Muse is available on iOS, Android, and muse.ai, with support for Meta’s AI glasses coming soon.

Pricing

Meta is offering three tiers:

• Free: Core agent functionality with limited usage
• $20 per month: Expanded access and faster response
• $100 per month: Full capability, priority access

This is Meta’s first paid AI consumer product. The company has historically made money through advertising, so charging directly for AI is a meaningful strategic shift. It signals that Meta sees agent subscriptions as a real revenue line, not a marketing exercise.

What This Means for Business

Consumers will expect agents to complete tasks, not just suggest them. Muse normalises autonomous AI at scale. When hundreds of millions of people use an agent to shop, book, and schedule, they’ll arrive at every business interaction with a higher baseline expectation. If your customer experience still requires a human to navigate your site manually, that gap becomes more obvious and more costly.

Agentic commerce is the next frontier. Muse can browse and buy on behalf of users. That means your product listings, checkout flows, and data quality matter more than ever. An agent evaluating products on a user’s behalf will behave differently to a human browsing. Businesses that think about how their products appear to an AI intermediary, not just a human eye, will have a structural advantage.

The consumer layer just caught up with enterprise AI. For the last two years, enterprise adoption of AI agents has been ahead of consumer behaviour. Muse closes that gap. Employees who use personal AI agents at home will increasingly expect the same at work. If your business hasn’t started building its own agent infrastructure, the expectation gap is only going to grow.

Small businesses and professional services need to pay attention. The $20 tier means Muse isn’t just for tech-savvy early adopters. Anyone with a smartphone and $20 can now delegate tasks to an AI. For accountants, consultants, recruiters, and anyone else who currently fills inbox-driven, appointment-driven workflows, consumer AI agents are a reshaping force.

The Bigger Pattern

Muse is part of a broader shift in how AI reaches people. Earlier this year, Google expanded its Gemini agents, Microsoft deepened Copilot integration across Windows, and Salesforce’s Agentforce went mainstream in enterprise deployments. The consumer layer has been the missing piece.

That piece just arrived.

Businesses building their own AI agent capabilities through platforms like Omni Ops or Omni Apps are ahead of the operational learning curve. Consumer products like Muse drive up the floor of what customers, employees, and partners expect from any digital interaction. The question is not whether to build agent capabilities but how fast you can do it before the gap between your AI maturity and your customers’ expectations becomes a competitive liability.

Muse’s US launch is today. The international rollout timeline hasn’t been confirmed. Meta will move fast.