A glowing AI assistant connects cloud storage, analytics, code, security, and team chat in a modern office.
Meta has decided that ads, messaging apps and a consumer AI agent aren't enough. Its next target is the corporate IT budget. On September 28, 2026, the company launched Meta Enterprise Platform, a new business unit that puts Meta in direct competition with Microsoft, OpenAI and Anthropic for enterprise AI agent spending. Meta's newsroom post introducing the unit opens by saying "Today we are starting Meta Enterprise Platform to help businesses use AI to grow, and Chirantan "CJ" Desai will join Meta as Chief Enterprise Platform Officer".

For Windows admins and Microsoft 365 shops, the point is simple. Microsoft now has one more company competing to be your agent platform. What Meta has published so far is mostly a business plan, though. It hasn't published product specs.

What Meta actually announced​

Mark Zuckerberg's statement names the parts. Meta says the unit will use our strengths that few other companies have: advanced models, leading agents, large-scale infrastructure, and years of working closely with many businesses. The first offering is a bundle of products Meta already sells: "Initially, we will focus on bringing our full technology stack, including the Muse agent, Meta Business Agent, Muse API, Muse Code, and more to businesses and developers to help them grow."

Zuckerberg's post on X gives the move a lot of weight. It says Meta is starting the next major pillar of our business. That's a big label for a company whose revenue has come almost entirely from advertising.

Key points:

  • New unit: Meta Enterprise Platform, which Meta describes as its next major business pillar.
  • Initial stack: Muse agent, Meta Business Agent, Muse API, Muse Code, "and more."
  • Leader: CJ Desai, Chief Enterprise Platform Officer, reporting directly to Zuckerberg.
  • Not yet published: pricing, deployment options, supported clouds, identity integration, admin controls, compliance certifications, data-residency terms and a customer availability date.

The hire: MongoDB's CEO leaves for Menlo Park​

The leadership move got the most attention from Wall Street. Meta says Desai joins us from MongoDB, where he was CEO and President. Before that he led product and engineering at Cloudflare and spent nearly eight years at ServiceNow, including as President and COO.

MongoDB confirmed the change in its own press release the same morning. It said Desai had stepped down as president and CEO effective immediately to take a senior role at Meta Platforms. Dev Ittycheria, who ran MongoDB from 2014 to 2025, returns as interim president and CEO. MongoDB says its board has started a search for a permanent CEO and that it is reaffirming its Q3 and full fiscal-year 2027 guidance. The timing is awkward because MongoDB's Investor Day at Nasdaq MarketSite is scheduled for September 29. Constellation Research also noted that the Meta news landed a day before MongoDB's investor day in New York.

CNBC reported that Desai, who was in his role at MongoDB for less than a year, will be leading the social media giant's new Meta Enterprise Platform for businesses. Reports of MongoDB's share-price drop vary depending on when they were written. NewsBytes said the stock fell 14% in premarket trading. SEE News later reported shares falling around 24% following the news. Treat any single figure as a snapshot of a moving target.

Desai's resume matters for this job. ServiceNow sells IT workflow software, Cloudflare sells network and security infrastructure, and MongoDB sells a developer database. All three sell to CIOs and developers, which is the audience Meta has never really had.

Desai's pitch, and what it doesn't cover​

In his statement, Desai argues that Meta has a unique role to play because it is bringing together advanced models and leading agents with a proven track record of helping millions of advertisers and hundreds of millions of businesses scale. He adds that the unit will focus on turning its AI stack into products and services that companies can deploy for their own businesses. As with Muse, security and privacy are built into Meta's enterprise products from the outset.

That is a stated aim. It isn't an architecture document. Any security team would want answers to these questions before a pilot:

  • Where does tenant data live, and who holds the encryption keys?
  • Does it support SSO and conditional access through existing identity providers such as Microsoft Entra ID?
  • What audit logging exists, and can it feed an existing SIEM?
  • Which compliance certifications are in hand, not just planned?
  • Is customer data excluded from model training by contract, not only by policy?

Meta hasn't answered any of these yet. Meta's consumer Muse documentation does describe safeguards such as approval prompts before sensitive actions and an audit trail. Those descriptions apply to a personal agent. Nobody should assume the same terms or implementation will carry over to an enterprise contract.

What's already in the box​

The product names in the bundle aren't vaporware. Meta has shipped them over the past several months.

Muse Code is the part developers are most likely to try first. Meta AI Research launched it in beta on August 5, 2026 as a terminal coding agent powered by Muse Spark 1.2. Meta says it can plan, write and validate changes across large repositories while coordinating persistent background subagents. Its runtime writes every model call, tool run, approval and edit to a local event log, so a session can resume where it stopped after a crash. A developer post dated August 31 said Muse Code was out of beta. That post added inter-session messaging, multi-agent "workflow" orchestration, a rewind feature (double Esc), a developer preview of a TypeScript SDK and monthly subscription plans.

Muse Spark 1.3 arrived September 2. Meta says it improves agentic and coding performance and is available through Muse Code and the Meta Model API. Meta also claims that in comparisons by its own engineers, the model used about 20% fewer tool calls and about 25% fewer tokens than 1.2. Those are vendor measurements, not independent benchmarks.

Meta Business Agent debuted in June as agent infrastructure for customer conversations on Meta's messaging apps. Muse itself is Meta's personal AI agent, introduced September 8. CNBC reports it has seen a rush of consumer demand, overtaking OpenAI's ChatGPT as the leading free iOS app.

A practical Windows note​

The Muse Code install instructions in Meta's August 5 and August 31 posts cover macOS or Linux only, using a shell one-liner. Meta doesn't list native Windows support in those announcements. Windows-based development teams evaluating the enterprise bundle should ask Meta directly about Windows support rather than assume it. Meta's posts don't say whether a Linux environment such as WSL is a supported option.

How it compares with Microsoft and OpenAI​

Neowin summarizes Microsoft's position as a stack under the "Microsoft IQ" framework, built around Microsoft 365 Copilot, Copilot Studio, Azure and Agent 365. In that account, Agent 365 serves as a control plane where organizations discover, govern, secure and monitor AI agents. Neowin describes OpenAI's Frontier as a platform that connects agents to data warehouses, CRM systems and internal applications, with execution, governance, identity controls and monitoring built in.

MicrosoftOpenAIMeta (as announced)
Enterprise agent platformCopilot Studio, Agent 365 control planeFrontierMeta Enterprise Platform
Productivity/identity footprintMicrosoft 365, EntraVia connectorsNone comparable announced
Developer coding agentGitHub CopilotCodexMuse Code
Customer-messaging reachTeams (internal)Via partnersWhatsApp, Messenger, Instagram
Governance details publishedYesYesNot yet

My analysis: Microsoft's biggest advantage is that it already owns the identity, the documents and the admin console. An agent that can't be governed in Entra and Purview has a harder sell in a regulated Microsoft 365 tenant. Meta's advantage is the reverse. It already sits between businesses and billions of customers on WhatsApp and Messenger, and Microsoft has no equivalent. So Meta's early wins may be customer-facing agents rather than internal copilots.

The skeptic's corner​

Is a consumer company pivoting to enterprise a punchline? Not necessarily. Constellation Research points out that Amazon launched Amazon Web Services and Google launched Google Cloud to chuckles only to become a major hyperscaler. Motive matters too. CNBC notes that investors have demanded a payoff from the company's hefty AI investments on new infrastructure and tools. Enterprise revenue is one way to get it.

The counterpoints are real, though:

  • Trust: Meta's privacy history will get extra attention in procurement reviews, fairly or not.
  • Enterprise muscle: Selling to CIOs requires support organizations, SLAs and partner channels that take years to build. That's presumably why Meta hired Desai.
  • Specifics: Until Meta publishes pricing, controls and certifications, the platform can't be evaluated.

Bottom line for IT pros​

Meta Enterprise Platform is confirmed and has an experienced enterprise leader. It starts with products that already work, including a capable coding agent and a large messaging footprint. For now it is an organizational launch, and the product details IT buyers need haven't been published. Microsoft 365 shops don't need to change anything yet. It's worth watching, especially if customer engagement runs through WhatsApp. When Meta's sales team calls, have the questions above ready.


Update: Additional details (September 30, 2026)​

MarketWatch reports that Meta Business Agent is already used by more than 1 million businesses across WhatsApp and Messenger. Meta has not yet said how many of those businesses use advanced agent workflows rather than basic messaging tools.

The report also notes that Meta shut down Workplace in May, after announcing a phased wind-down two years earlier. That history gives IT buyers another support-lifecycle question to raise before committing business processes to Meta’s new platform.

 

References

  1. Meta takes on Microsoft and OpenAI with its new enterprise platform for AI agents - Neowin Neowin 2026-09-28T13:46:01+00:00
  2. Meta is pitching a bold new identity - and the payoff could be big Longbridge 2026-09-30T12:28:32+00:00
  3. Meta hires MongoDB CEO CJ Desai to lead enterprise unit. MongoDB shares crater cnbc.com