On September 28 (US time), Meta announced it is launching Meta Enterprise Platform, a new enterprise AI business. The plan is to bring the personal agent Muse, the customer-facing Meta Business Agent, and the developer-focused Muse API and Muse Code to companies. Meta said it is bringing in Chirantan "CJ" Desai, formerly president and CEO of MongoDB, to lead the effort, reporting directly to CEO Mark Zuckerberg. The customer-facing Meta Business Agent and the personal Muse agent had both been announced earlier. What has changed is that Meta now says it will treat them as a single enterprise business and has assigned a dedicated executive to it.
How is Meta repackaging what already exists?
Meta's enterprise AI did not appear out of nowhere on September 28. In June, the company announced Meta Business Agent, which answers customer questions on WhatsApp and Messenger. Meta described it as a tool for product recommendations, bookings, and qualifying leads, and it also introduced the Meta Business Agent Platform, where companies build, tune, and run their agents. For larger enterprises, Meta said it would offer administrative controls for setting agent behavior rules, along with performance measurement. It also cited connections to hundreds of systems, including Shopify and Zendesk. However, these claims come from Meta's own product description, not from independent verification that every integration works the same way.
In the June explanation, businesses could start using the customer-facing agent for free, with paid plans planned for the future. Features that help businesses themselves, such as summarizing unread customer messages each morning, were to be offered first to select companies. In other words, the scope of availability at the time differed between automated replies to customers and support for business operations.
Meta said at the time that more than 1 million businesses were using Meta Business Agent on WhatsApp and Messenger. That figure is Meta's own claim about an existing customer-service product. It does not represent the number of contracted customers or deployments for the new Meta Enterprise Platform. The similarly named "Meta Business Agent Platform" and "Meta Enterprise Platform" are also distinguished in the announcement: the former is the foundation for building and running customer-facing agents, while the latter is the name of the new business delivering a broader product lineup to enterprises.
Paid plans are not new either. In "Meta One," announced on September 15, the Essential tier for businesses and creators includes a usage allowance for Meta Business Agent on WhatsApp, with prices starting at $14.99 per month. Advanced starts at $49.99 per month and includes more agent response capacity. Expert starts at $149 and Max at $499. Meta also states that pricing and available features vary by region, app, and account. These are Meta One prices, not a price list for the entire Meta Enterprise Platform.
Muse, announced on September 8, is a personal agent that works on the web and in connected apps on a user's behalf. It is rolling out in phases in the US on iOS, Android, and the web, and Meta says it runs in a dedicated virtual machine for each user, called Muse Secure VM. The latest announcement names this consumer product as an enterprise candidate as well. But the details on the personal Muse's availability regions and virtual machine cannot simply be read as enterprise contract terms or data protection conditions.
Seen this way, Meta is trying to extend beyond its existing business of automating customer conversations to cover employee work and developer use. The September 28 official announcement lists four product names as the "initial" scope, but it does not say when they will be unified under a common admin console, contract, or data foundation. What can be confirmed for now is the intention to sell this product family to enterprises.
What does a CEO-level reporting line signal?
Meta announced that Desai will become head of the enterprise platform and report directly to Zuckerberg. According to Meta, before MongoDB, Desai led product and engineering at Cloudflare and served in roles including president and COO at ServiceNow. Meta's explanation of the hire is that it has chosen someone with experience embedding products in enterprise operations and expanding customer contracts, not just conducting model research. But a title and résumé do not reveal the new business's revenue scale or product timelines.
The move also brought an immediate change at MongoDB. According to MongoDB's September 28 announcement, Desai stepped down as president and CEO effective the same day, and former CEO Dev Ittycheria became interim president and CEO. The board will search for a permanent successor. It is easy to mix up the two men here: Desai is the one moving to Meta, while Ittycheria is returning to MongoDB as its top executive.
Placing Desai under the CEO suggests Meta has made selling the AI it has offered to advertisers and messaging-app businesses as enterprise software a management priority. However, Meta has not announced how contracts will change for companies moving from the existing Meta Business Agent to the new business, or how products such as Muse Code can be purchased in combination.
What enterprises still can't determine
While the September 28 announcement lists four products including Muse, it does not give the platform's own pricing, launch date, target regions, or terms for handling enterprise data. This check covers the full text of the official announcement of the same date; it does not mean such terms do not exist on other product pages or in future individual contracts. Desai said security and privacy will be built in from the design stage of enterprise products, but the announcement contains no specific terms that companies could check on data retention periods, administrator permissions, audit logs, or whether data is used for training.
This is where purchasing decisions hinge. Meta One's monthly pricing lets small businesses estimate costs for some of the existing customer-service features. But for plans that involve using Muse for internal work, or having developers use the APIs and Muse Code, the product names alone are not enough to set a budget or a management approach. Nor does the US launch of the personal Muse mean Japanese companies can adopt it on the same terms.
Meta is trying to tie customer service, a personal agent, and developer tools into one enterprise business. Whether that vision becomes a product enterprises can actually buy will depend on the scope and pricing of each product, and on whether Meta can show who manages enterprise data and which operations can be audited.