Meta introduced Muse on September 8 as a personal AI agent, initially launching it in the U.S. across iOS, Android and muse.ai. The service expanded to Canada later that month. The company says Muse can use a browser, fill out forms, connect to apps, keep working after the app is closed and ask for approval before sensitive actions such as sending emails or making purchases.
Early download estimates point to attention, not settled adoption
TechCrunch reported on September 25 that market-intelligence estimates placed Muse at 3.4 million downloads according to Sensor Tower, 4.3 million according to Apptopia and roughly 2.3 million according to Appfigures. Those early estimates varied enough that they should be treated as directional rather than an exact user count.
The stronger signal is that Muse rose quickly in U.S. app-store rankings during the launch window. TechCrunch reported that Muse reached the top of the U.S. App Store on September 18 and the top spot on Google Play on September 19. That momentum followed Meta’s broader push at Connect 2026, where the company promoted smart-glasses integration, more connectors and an email address for the agent.
Meta is already pushing Muse into business workflows
On September 29, Meta announced Muse for Small Business, adding connectors for tools including Asana, Box, Canva, Dropbox, Figma, Intuit QuickBooks, Klaviyo, Notion, Shopify, Slack, Stripe and Zoom, along with Facebook and Instagram business accounts.
Meta’s pitch is that Muse can analyze sales, campaigns and social activity, prepare growth plans, draft customer communications, watch expenses and help create ads. The company says nothing publishes, sends or spends without approval.
This changes the adoption question. A casual AI app can survive on curiosity and occasional use. A business agent has to earn permission to read customer records, books, campaign data, payment information and messages. That is a higher-friction decision, especially for small businesses without dedicated IT or security staff.
Small businesses may be a more natural fit than households
Axios argued on September 30 that AI agents still face a “normal-people” adoption problem because much of daily life happens away from screens. A December 2025 Gallup report, based on an October 2025 U.S. survey, also found that using AI tools was one of only two activities that failed to earn at least “somewhat important” ratings from a majority of adults in their view of a good life.
Workplace adoption data points in a similar direction. The St. Louis Fed reported in September 2026 that AI use was highest in computing and professional jobs and lowest in hands-on or face-to-face work. Its analysis found adoption above 50% in only 40% of occupations, while fewer than 3% of job tasks had adoption rates above 50%.
Small businesses sit closer to Meta’s strongest opening. Their work often already runs through digital systems: Facebook Pages, Instagram accounts, ads, ecommerce platforms, bookkeeping software, calendars, email, files and chat. The value of an agent rises when it can move across those systems instead of answering in a single chat window.
Trust is the product
Meta is trying to answer the trust problem by saying Muse runs in a dedicated cloud virtual machine, stores credentials separately, uses a separate Sentinel system to approve internet-facing actions, provides audit trails and does not share Muse conversations or virtual-machine data with Meta’s ad systems.
Meta also says users can choose connector permissions, disconnect services and opt out of having interactions used to train AI models. The company has said it plans to introduce Muse Confidential VM later this year, with the virtual machine encrypted using a key held by the user.
Those safeguards are central to adoption because agentic systems need access before they can be useful. The same access that lets Muse draft a customer reply or flag an unusual expense also makes errors, over-permissioning and prompt-injection attacks more consequential than with a standard chatbot.
Meta’s adoption strategy is bigger than one app
On September 28, Meta announced Meta Enterprise Platform, a business unit built around AI products and services. Meta said the initial focus would include the Muse agent, Meta Business Agent, Muse API, Muse Code and related infrastructure.
The company also said former MongoDB CEO Chirantan “CJ” Desai would join as chief enterprise platform officer, reporting to Mark Zuckerberg.
Taken together, the September 28 enterprise announcement, the September 29 small-business rollout and the Connect glasses integration show Meta attempting to make Muse a permission layer across consumer, small-business and enterprise workflows.
The next adoption signals to watch
App-store rankings are easy to promote, but they are not the decisive measure. More revealing signals will include repeat use after the launch surge, how many users connect sensitive services, how many businesses pay for subscription plans and whether Meta can show reliable outcomes without major privacy or safety failures.
Get new small business insights by email
Practical ideas and useful articles to help you make better business decisions.
Tech Help Canada Staff researches, writes, and reviews practical content for business owners and professionals. Our coverage spans business, marketing, SEO, technology, and the tools and systems people use to grow and operate online. We focus on clear, useful information backed by research, hands-on experience, and editorial review. Learn more about our team and editorial standards. Need help with something? Contact Us
