Meta expands Muse AI agent to small businesses and enterprise software

By Splitfeed EditorialLast updated Sep 30, 2026, 19:50 UTC

On September 29, 2026, Meta launched Muse for Small Business, an expansion of its Muse AI agent designed to move the product from consumer use into small business workflows and enterprise software. The new offering includes a subscription model and third-party app integrations, though the specific platforms connected to the service were not named in available reporting. A central feature of the expansion is that Muse can execute business tasks autonomously, but only after the business owner gives explicit approval for each action. Mark Zuckerberg framed the launch as part of a deliberate effort to diversify Meta's revenue base beyond its historically dominant advertising business.

The launch matters beyond the product itself for a few reasons. Meta is positioning agentic AI, meaning software that takes real-world actions on a user's behalf, as a commercial growth engine. The subscription model signals that Meta wants recurring revenue from business customers rather than relying solely on ad impressions. What drew immediate scrutiny, however, is a term buried in Meta's own documentation: cardholders bear full financial responsibility for any purchases Muse makes. That clause, combined with reported privacy concerns and alleged security flaws in the product, has created a genuine dispute about whether the approval-gating mechanism provides adequate protection, or whether it is too easy to bypass or misuse in a business context.

No published reactions from Pro-AI, Anti-AI, or Middle Ground commentators were available at the time of writing. Typically, a Pro-AI camp would highlight the productivity gains for small businesses that lack dedicated staff for routine tasks, pointing to the approval requirement as a reasonable safeguard. An Anti-AI camp would be expected to focus sharply on the financial liability clause and the privacy and security concerns, arguing that routing business transactions through a corporate AI agent creates unacceptable risk for small operators. A Middle Ground position would likely accept the premise of AI business tools while calling for clearer liability frameworks and independent security audits before wider adoption.

The arguments that will shape this story going forward center on whether Meta addresses the liability and security criticism in its terms or product design, and whether any of the reported privacy issues result in regulatory attention or formal complaints. The specific third-party integrations Meta chooses to announce or certify will also be a signal of how seriously enterprise partners are taking the security concerns raised at launch.