Meta CEO Mark Zuckerberg and Meta President and Vice Chair Dina Powell McCormick attend the annual Allen and Co. Sun Valley Media and Technology Conference at the Sun Valley Resort in Sun Valley, Idaho, U.S., July 9, 2026.
Brendan McDermid | Reuters
Meta introduced its AI personal agent app on Tuesday and is asking a subset of users to pay at a time when the company and broader artificial intelligence industry face intensifying scrutiny and public skepticism.
The app, which was given the internal code name Hatch, is powered by the Muse Spark family of foundation models that Meta AI chief Alexandr Wang has been debuting at a breakneck pace since April.
Meta said the Muse app lets people offload digital tasks like booking appointments, filling out electronic forms and even monitoring home security camera feeds to AI-powered assistants. The company designed the app so “it feels very approachable and friendly and explainable, and it doesn’t feel too complicated,” Wang told CNBC in an interview.
“Behind the scenes, Muse might be doing very advanced coding workflows, or building sophisticated integrations, or doing quite a lot of heavy lifting while keeping that very sort of simple for the user,” Wang said.
Wang said the Muse personal agent will be available in either a free tier or through monthly subscription plans of $20 or $100, depending on usage.
Meta CEO Mark Zuckerberg, who hired Wang last year as part of a $14 billion investment in his startup, Scale AI, has been dropping hints to investors that the company’s next big AI bet involves supercharged digital assistants that sift through emails or look for deals on furniture.
Zuckerberg has said that supercharged digital assistants represent the next big leap for AI models, pitching them as one of the reasons that Meta is spending heavily on data centers and infrastructure. In the company’s earnings call in July, Zuckerberg said “new personal agents that will be the foundation for our next wave of products and revenue lines in the months and years ahead.”
Meta’s launch of its personal assistant comes amid a particularly fraught time for the company. Meta recently agreed to pay nearly $17 billion in a major settlement with a coalition of state attorneys general that had sued the company for misrepresenting the prevalence of harm on apps like Facebook and Instagram. Meta still faces a wave of personal injury and school district lawsuits involving similar allegations.
Meanwhile, there are rising concerns about the potential cybersecurity risks posed by AI agents and their underlying foundation models, and the country faces a growing backlash against AI data centers and the companies that are building them.
But Meta is also under pressure from Wall Street to show returns on its AI investments, as the company tries to diversify its business and reduce its reliance on digital ads. The personal agent follows other newer initiatives like the Muse Code AI agent for developers and various subscription plans that Meta has recently rolled out to generate new lines of business.
In addressing security concerns, Wang said the app runs within “its own isolated environment” inside Meta’s computing infrastructure, and “never sees your actual passwords or payment details and asks before doing anything sensitive.”
Users must opt out of letting Meta use their interactions with the Muse personal agent to train the company’s AI models. If they don’t, the company will scrub “critical personally identifying information” before using the agent conversations and interactions to improve AI models, said David Singleton, Meta vice president of engineering.
The company is exploring potential monetary initiatives, such as taking a cut of AI agent-related shopping transactions, but hasn’t settled on any concrete plan, Wang said.
“We think the commerce business model is potentially really interesting for this product because of how much it enables people to to actually find the things that they are excited about and ultimately fulfill a lot of those purchases,” Wang said.
Users of the Muse personal agent app will be able to see a feed that shows them various updates from their connected Facebook and Instagram accounts, or certain articles or stories gleaned from the web, depending on what they want the feed to aggregate and summarize. The agent will also be available via WhatsApp, but that version won’t contain the app features like the feed or an ideas tool that recommends AI agent trips and tricks, Wang said.
“The core functionality that you have in the main conversation with your agent is the same across those two services,” Wang said.
Meta is playing into a major industry trend. Since the rapid rise of OpenClaw, used by coders to manage the AI models powering digital assistants, companies like OpenAI and Google have introduced similar agentic tools and features alongside startups like Town.
Wang said the Muse personal agent is intended to be “more accessible to the broader audience,” compared to competing products, but conceded that it’s still “pretty early in in this new era of personal agents.”
Meta is asking third-party security researchers to find vulnerabilities within the Muse personal agent through a so-called bug-bounty program that includes financial rewards if they discover certain issues.
“We’ve hardened Muse based on extensive dogfooding, agentic red teaming, and against issues found in real adversarial scenarios by security researchers in our private bug bounty program,” the company said in a technical blog post.
Meta said the personal agent will be available to U.S. consumers via iOS, Android or through a standalone website. It will eventually be accessible through the company’s Ray-Ban Meta glasses.
WATCH: Colorado AG says Meta lawsuit’s most important takeaway is the change in behavior, not money.
