Meta Launches Enterprise AI Platform and Hires MongoDB CEO CJ Desai to Run It

Meta Platforms is opening a new enterprise AI business and has hired MongoDB President and CEO Chirantan “CJ” Desai to run it, extending the company’s AI push beyond consumer products, advertising tools and developer models into software and services aimed directly at businesses.

The company calls the new operation Meta Enterprise Platform. Mark Zuckerberg said the unit will bring together Meta’s AI models, agents, infrastructure and experience serving businesses, with an initial focus on products including Muse, Meta Business Agent, Muse API and Muse Code. Desai will become Chief Enterprise Platform Officer and report directly to Zuckerberg, according to Meta’s Sept. 28 announcement.

Meta did not disclose a start date for Desai, broad pricing for the new platform or a detailed rollout schedule. It also did not say who will succeed him at MongoDB. The announcement instead establishes the leadership and product direction for a business Meta is positioning as another major part of the company alongside its existing consumer, advertising and hardware operations.

Meta is grouping its AI products into an enterprise offering

The new platform is not starting from a blank slate. Several of the products Meta named already exist in consumer, developer or business-facing forms, and the enterprise effort appears designed to package those capabilities for organizations that want to deploy Meta’s AI stack in their own operations.

Muse is Meta’s personal AI agent, introduced earlier in September. The company says it can work across connected services on a user’s behalf, including handling tasks such as email and travel booking, while running inside a dedicated Muse Secure VM. Meta has emphasized security, permission controls and auditability in the product, including requiring user approval before certain sensitive actions.

Meta Business Agent is already aimed at companies. When Meta expanded the product in June, it said more than one million businesses were using a Business Agent on WhatsApp and Messenger and that more than one billion active threads with businesses take place each day across WhatsApp, Messenger and Instagram. Meta also introduced a Business Agent Platform for larger organizations to build, customize and deploy agents with enterprise controls and connections to outside systems.

The developer side includes Muse Code and access to Muse models through Meta’s model API. Meta describes Muse Code as a multi-agent coding tool built around its Muse Spark models, while its model API gives developers direct access to those models for their own applications. Under the new enterprise unit, Meta says these products and related technology will be offered as part of a broader stack for businesses and developers rather than remaining separate product lines with no common enterprise umbrella.

What Meta has not yet spelled out is how those pieces will be sold together. The company has not announced a platform-wide subscription structure, a general-availability date for the full enterprise offering or a detailed list of customer commitments. Those details will determine whether Meta Enterprise Platform develops mainly as a bundle of existing tools or becomes a more distinct enterprise software business.

Desai brings enterprise software experience after less than a year as MongoDB CEO

Desai’s appointment gives the new unit a leader whose recent career has been concentrated in enterprise software, cloud infrastructure and AI products. MongoDB appointed him President and CEO effective Nov. 10, 2025, succeeding longtime chief executive Dev Ittycheria. His move to Meta therefore comes less than a year after taking MongoDB’s top job.

Before MongoDB, Desai was President of Product and Engineering at Cloudflare. He previously spent nearly eight years at ServiceNow, where his roles included President and Chief Operating Officer. Earlier positions included executive roles at EMC and Symantec, after beginning his career at Oracle.

That background is closely aligned with what Meta says it wants from the new business. Zuckerberg cited Desai’s experience in AI, infrastructure, business applications and security, while Desai said the platform will focus on turning Meta’s AI stack into products and services that companies can deploy for their own businesses. He also said security and privacy are intended to be built into those enterprise products from the outset.

Desai’s departure also creates a fresh leadership question for MongoDB. The database company had been emphasizing enterprise AI under his tenure, including new capabilities for agent memory, embeddings, search and private-cloud deployments. Meta’s announcement does not provide information about MongoDB’s succession process, and MongoDB had not announced a replacement in the primary company materials inspected for this story.

The move gives Meta a more direct route into enterprise AI spending

Meta already has several points of contact with businesses through advertising, messaging, commerce and developer tools. The Enterprise Platform broadens that relationship by targeting organizations that want AI agents and model-powered software for internal workflows, customer service, software development and other operating tasks.

The distinction matters because Meta’s biggest established businesses have historically been built around consumer services and advertising rather than the enterprise software model used by companies such as Microsoft, Salesforce, ServiceNow or MongoDB. Meta is not presenting the new unit as a replacement for its advertising business. Instead, Zuckerberg described it as an additional pillar that can use the company’s models, infrastructure and business relationships to sell a different class of AI products.

There is already some overlap between the two sides. Meta Business Agent operates inside WhatsApp, Messenger and Instagram, placing enterprise AI tools directly in channels where companies communicate with customers. Muse Code and the model API address developers, giving the enterprise unit a route into technical teams building their own software. Muse provides another set of agent capabilities that Meta says can be adapted for organizations.

The Sept. 28 announcement leaves the commercial model largely open. Meta named the initial technology stack and the executive who will lead it, but did not disclose revenue targets, expected investment, customer contracts or when the full enterprise platform will become broadly available. Desai’s start date and MongoDB’s leadership response are also still to be specified in the primary announcements reviewed for this story.

Those omissions make the launch an organizational and product-direction announcement rather than a fully detailed commercial rollout. The next concrete disclosures to watch are Meta’s product availability, pricing and customer plans for the Enterprise Platform, along with MongoDB’s decision on leadership after Desai’s move.