Meta Launches Enterprise AI Platform, Taps MongoDB CEO

Meta wants to sell the AI it built for billions of consumers to the enterprise, and it hired a sitting public-company CEO to do it.

On Monday, Meta launched Meta Enterprise Platform, a new business unit that CEO Mark Zuckerberg described as "the next major pillar of our business." The unit will be led by Chirantan "CJ" Desai, who stepped down as president and CEO of MongoDB the same day to become Meta's chief enterprise platform officer, reporting directly to Zuckerberg.

What Meta Is Bringing to Enterprises

Zuckerberg said the platform will initially focus on bringing Meta's full technology stack to businesses and developers. He named four components:

Component	What It Is
Muse agent	Meta's personal AI agent, which runs on a dedicated cloud virtual machine and can browse, fill out forms and complete tasks on a user's behalf
Meta Business Agent	A customer-facing agent for businesses that Meta introduced in June
Muse API	Developer access to Meta's Muse models, building on the Meta Model API preview Meta opened when it shipped Muse Spark 1.1
Muse Code	A terminal-based coding agent that launched in beta in August, powered by Muse Spark 1.2

Zuckerberg said the platform draws on "strengths that few other companies" have, like advanced models, agents, large-scale infrastructure and years of working with businesses. Meta did not share pricing, launch timelines or which customer segments it will target first.

Why Meta Tapped Desai

Zuckerberg pointed to Desai's record building full-stack software across AI, infrastructure, business applications and security, as well as his industry network. Before MongoDB, Desai led product and engineering at Cloudflare and spent nearly eight years at ServiceNow, including as president and COO.

Desai said the platform will focus on "services that companies can deploy for their own businesses." He added that security and privacy are built into Meta's enterprise products from the outset, as with Muse. "Our goal is to make Meta the place enterprises come to scale their businesses."

The departure left MongoDB without a permanent CEO less than a year after Desai took the role. Desai had taken over from Dev Ittycheria in November 2025. Ittycheria, who ran the company from 2014 to 2025, will now step back into the role as interim president and CEO. 

Investors did not take the departure news calmly, with MongoDB shares falling sharply on Monday.

Meta's Enterprise Credibility Test

Meta already works with hundreds of millions of businesses through its advertising and messaging tools. Selling agents that act inside a company's systems is a different sale, one that runs through security reviews, procurement and compliance teams.

Several safeguards in the consumer version of Muse line up with what enterprise security teams typically require. Muse runs on its own isolated cloud computer, and a separate Sentinel agent must approve anything Muse sends to the internet. The agent can't see stored passwords or payment methods, asks before sensitive actions and keeps a full audit trail. Meta also plans to release Muse Confidential VM later this year, encrypting the entire environment with a key only the user holds.

Muse's debut also raised a practical question for any company deploying agents: What happens when a website doesn't want your agent there? Within weeks of launch, Amazon blocked Muse from shopping on its platform, claiming that:

1. Meta never notified Amazon
2. Muse doesn't identify itself when browsing
3. Muse appears to capture and store customer credentials

If AI agents can be shut out of Amazon, they can be shut out of a supplier portal. Companies that plan to deploy AI agents outside their own systems will need to settle how those agents identify themselves and obtain permission. 

Meta also enters a crowded field. OpenAI recently gave developers a managed foundation for building AI agents, and Muse Code competes directly with established coding agents from OpenAI and Anthropic.

What enterprise leaders should watch:

• Packaging and pricing for Muse agents and Business Agent in enterprise contexts
• Data commitments, including whether enterprise data stays out of model training and Meta's ad systems, as Meta has promised for consumer Muse
• Partner strategy, given Zuckerberg's emphasis on Desai's industry network
• Model maturity, since Meta acknowledged gaps in coding and long-horizon agentic tasks when it unveiled Muse Spark in April