Meta’s new enterprise announcement names four AI products. The key detail is that several were already available before this week’s announcement.
On September 28, Meta announced Meta Enterprise Platform, an effort to bring its AI models, agents, and infrastructure to businesses and developers. Mark Zuckerberg named the Muse personal agent, Meta Business Agent, Muse API, and Muse Code as the initial focus. He also said Chirantan “CJ” Desai, who led MongoDB as CEO and president, is joining Meta as chief enterprise platform officer and will report to him. The announcement is brief: it does not include a launch schedule, a combined enterprise price, or contract terms for the new platform.
What is new, and what was already available?
The new item is Meta’s enterprise-platform initiative and the appointment of its leader. The product names are not all new. Meta introduced Business Agent in June and Muse Code in August. It introduced the Muse personal agent on September 8. Its developer documentation also describes an existing Meta Model API.
| Name in the announcement | What already exists | What the new statement does not settle |
|---|---|---|
| Muse agent | Meta introduced Muse as a personal agent on September 8 and said it was rolling out in the US. | How businesses would administer the agent, connect company data or buy enterprise access. |
| Meta Business Agent | Meta introduced Business Agent and its Business Agent Platform in June, with customer-service uses across its messaging apps. | Whether the new Enterprise Platform changes current deployment, support or commercial terms. |
| Muse API | Meta already documents and prices its developer offering under the name Meta Model API. | Whether “Muse API” in this announcement is the same offering, a renamed offering or a wider enterprise API package. |
| Muse Code | Meta released the coding agent in beta in August and said at Connect on September 24 that it was out of beta. | What, if any, additional enterprise administration, purchasing or support terms apply. |
The API naming deserves particular care. Zuckerberg’s statement says Muse API; the current developer site calls the product Meta Model API. Meta has not explained that relationship in the enterprise announcement. There is no basis to claim either a separate new API or a complete absence of API documentation. For product-level detail, see our earlier guides to Muse’s permissions, Muse Code’s agent handoff, and Model API pricing and data use.
Availability and pricing are not a blank page.
Our earlier version wrongly called the API a planned developer product and suggested Muse Code lacked a release date. Meta’s Model API page already publishes model prices, while its Muse Code documentation describes subscription and pay-as-you-go options. These prices apply to existing products, not the Meta Enterprise Platform as a whole.
Meta’s own availability labels also need context. Its September 24 developer recap says Meta Model API is generally available globally and Muse Code is out of beta. The Model API product page still says “Public preview.” That inconsistency is in Meta’s published materials; businesses should confirm the status, region, and terms for the exact service they intend to use.
What an enterprise buyer still needs to ask
Desai’s stated aim is for companies to use Meta’s AI stack inside their own operations. He also promises security and privacy from the start. Those statements do not replace a product-specific contract or technical documentation. Business Agent Platform already has Meta-described controls and integrations; the question is how those existing capabilities relate to the newly named platform.
- Identify the exact product. Ask whether the proposed service is Muse, Business Agent Platform, Meta Model API, Muse Code, or a new Enterprise Platform package. Get the product name in writing.
- Confirm the commercial terms. Check the current price and availability, then ask whether an enterprise contract changes billing, limits, support, or service commitments.
- Trace data and permissions. Document which company systems an agent can access, what actions require approval, where data is retained, and which administrators can review activity.
- Run a bounded pilot. Test one real workflow and measure completed tasks, human corrections, and operating cost. A product demo is not evidence that the proposed deployment fits your controls.
The practical reading
Meta has announced a new enterprise direction, not four first-time product launches. Muse, Business Agent, Muse Code and a model API already have public histories, but the relationship between those offerings and Meta Enterprise Platform needs more detail. For a buyer, that distinction changes the next step: evaluate the product that exists today, then ask what the new platform actually adds.
Check Meta’s original materials.
- Read Meta’s Enterprise Platform announcement.
- See Meta’s Business Agent and Business Agent Platform announcement.
- Check the September 24 developer recap for API and Muse Code availability.
- Review Meta Model API and Muse Code documentation.
Which enterprise details would you need before testing Meta’s stack in a real workflow?
