Meta Enterprise Platform is Meta’s new business unit for selling its AI to companies, announced on 28 September 2026 with an unusual hire at its head. Chirantan “CJ” Desai, who had been chief executive of the database company MongoDB for less than a year, resigned the same morning to become Meta’s Chief Enterprise Platform Officer, reporting directly to Mark Zuckerberg.
Zuckerberg called it “the next major pillar of our business”. MongoDB’s investors saw it differently: the company’s shares closed down 18.5% that day after trading as much as 26.9% lower, while Meta’s own stock fell 4.8%. MongoDB brought back its long-serving former chief executive, Dev Ittycheria, on an interim basis.
This article explains what Meta Enterprise Platform will sell, who CJ Desai is and why Meta wanted him, what his exit means for MongoDB, and how Meta’s move changes the enterprise AI market. It ends with the questions a business should ask before building on Meta’s AI stack.
Table of contents
- What Meta Enterprise Platform Is
- Inside Meta Business Agent
- A Timeline of Meta’s Push Into Business AI
- Who Is CJ Desai?
- What Desai’s Exit Means for MongoDB
- Why Meta Enterprise Platform Matters to the AI Market
- Security and Trust Questions for Meta Enterprise Platform
- Questions Meta Has Not Yet Answered
- What Businesses Should Ask Before Adopting Meta’s Enterprise AI
- What Meta Enterprise Platform Means for UK Businesses
- What Happens Next
- Meta Enterprise Platform FAQs
- References
What Meta Enterprise Platform Is
In a post on Meta’s newsroom, Zuckerberg framed the launch around the company’s superintelligence ambitions. “Meta already serves billions of people at scale and helps hundreds of millions of businesses reach customers,” he wrote. “Today we are starting the next major pillar of our business, Meta Enterprise Platform, to help businesses use AI to grow and transform in new ways as well.”
Four strengths, one stack
Zuckerberg said Meta Enterprise Platform would use “strengths that few other companies have: advanced models, leading agents, large-scale infrastructure, and years of working closely with many businesses.” Initially it will bring Meta’s “full technology stack, including the Muse agent, Meta Business Agent, Muse API, Muse Code, and more to businesses and developers.”
What is already on sale
None of those products is new. What is new is a single unit, with its own executive, to package and sell them to organisations. The table sets out what each piece does today, based on Meta’s announcements and developer documentation.
| Product | What it does | Status before the launch |
|---|---|---|
| Muse agent | Personal AI agent that performs tasks such as email and travel booking | Launched to consumers on 8 September 2026 |
| Meta Business Agent | Answers customers, recommends products, books appointments and qualifies leads on WhatsApp, Instagram and Messenger | Global since 3 June 2026 |
| Muse API | Developer access to the Muse Spark models | Muse Spark 1.3 at $1.25 in and $4.25 out per million tokens |
| Muse Code | Coding agent for developers | Subscription tiers since September |
The pitch is breadth: one supplier for customer messaging, a general-purpose agent, model access and developer tools. For businesses already spending on WhatsApp and Instagram advertising, Meta Enterprise Platform promises to extend an existing relationship rather than start a new one.
Inside Meta Business Agent
Of the four products, Meta Business Agent is the one most businesses can use today, and it shows how Meta Enterprise Platform is likely to be sold. It launched globally on 3 June 2026 after nearly two years of testing in markets including India and Mexico, TechCrunch reported.
What it does
The agent answers customer questions, recommends products, books appointments and qualifies sales leads, then hands a conversation to a person when needed. It runs in WhatsApp Business, Instagram direct messages, Messenger and Meta Business Suite. Meta was also testing daily briefings that summarise overnight chats for the business owner.
How it is priced
Small businesses get the agent through some tiers of the WhatsApp Business Premium subscription. Larger businesses using the WhatsApp Business Platform pay according to token usage rather than per message. Meta said it was building a way for large enterprises to create custom agents, with integrations for Shopify, Zendesk and Shopee in development. Those enterprise features are the natural first job for Meta Enterprise Platform.
A Timeline of Meta's Push Into Business AI
Meta Enterprise Platform is the latest step in a sequence that has moved quickly through 2026.
| Date | Step |
|---|---|
| 3 June 2026 | Meta Business Agent available globally on WhatsApp, Instagram and Messenger |
| 9 July 2026 | Muse Spark 1.1 released with Meta’s first paid developer API |
| 8 September 2026 | Muse personal agent launched to consumers |
| September 2026 | Muse Code coding agent offered on subscription tiers |
| 28 September 2026 | Meta Enterprise Platform launched; CJ Desai hired to lead it |
The pattern is consumer first, business second. Meta builds a product for its billions of users, proves it at scale, then packages it for companies. Meta Enterprise Platform formalises the second step and gives it a single owner.
Who Is CJ Desai?
Desai is an enterprise software operator rather than an AI researcher. Zuckerberg described him as “an experienced enterprise leader with a track record of building full-stack software and delivering results in AI, infrastructure, business applications, and security.”
| Role | Company | Period, as reported |
|---|---|---|
| Senior roles, including President and COO | ServiceNow | Nearly eight years |
| President of product and engineering | Cloudflare | About a year |
| President and CEO | MongoDB | November 2025 to 28 September 2026 |
| Chief Enterprise Platform Officer | Meta | From 28 September 2026 |
Why Meta wanted an enterprise operator
Meta knows how to sell advertising to millions of businesses. It has much less experience selling software to large enterprises, where deals involve procurement teams, security reviews, service-level agreements and data processing contracts. Desai has spent his career on that side of the market, at ServiceNow, which sells workflow software to large companies, and at MongoDB, whose database is used by around 75% of the Fortune 100. Meta Enterprise Platform needs exactly that experience if it is to win contracts rather than sign-ups.
What Desai says he will build
Desai’s statement focuses on productising. “Meta Enterprise Platform will focus on turning its AI stack into products and services that companies can deploy for their own businesses,” he said. “As with Muse, security and privacy are built into Meta’s enterprise products from the outset. Our goal is to make Meta the place enterprises come to scale their businesses.”
Why a database chief suits an AI platform
The hire makes more sense when you look at where enterprise AI projects struggle. Models are increasingly easy to buy. Connecting them safely to a company’s own records, permissions and workflows is the hard part. MongoDB now describes itself as “the intelligent data platform for the AI era”, combining operational data, search, analytics and retrieval with embedding and reranking models from its Voyage AI unit. Desai has spent the past year selling exactly that connective layer. Meta Enterprise Platform will need it if Muse agents are to work with a customer’s data rather than only with public information.
What Desai's Exit Means for MongoDB
The news broke before US markets opened: Meta’s newsroom post went up at 8:36 a.m. Eastern. Desai stepped down “effective immediately”, and Ittycheria, who ran MongoDB from 2014 to 2025, returned as interim President and CEO while the board searches for a permanent successor with an executive search firm.
How far the shares fell
Reports of the fall ranged from “more than 17%” to 25%, because they were written at different times of day. The chart uses Yahoo Finance daily prices and compares each figure with MongoDB’s close of $410.44 on Friday 25 September. Each figure is (price ÷ $410.44) − 1.
About 17.7 million MongoDB shares changed hands, around 12 times the average of the previous seven sessions. CNBC noted the stock had already lost nearly a quarter of its value in 2026 before the announcement.
A business still growing
MongoDB was keen to stress that the business itself was not the problem. “MongoDB’s business is strong,” said chairman Tom Killalea. The company reaffirmed the guidance it gave on 1 September, when it reported second-quarter revenue of $771.8 million, up 30% year on year, and raised its full-year forecast to between $2.99 billion and $3.03 billion. It is holding an Investor Day at Nasdaq MarketSite on 29 September, scheduled before the news broke.
| MongoDB fact | Figure | Source |
|---|---|---|
| Q2 fiscal 2027 revenue | $771.8m, up 30% | Results, 1 September |
| Full-year revenue guidance | $2.99bn to $3.03bn, reaffirmed | Release, 28 September |
| Customers | More than 70,000 | Release, 28 September |
| Revenue growth under Ittycheria | About $35m to more than $2.3bn a year | Release, 28 September |
What MongoDB customers should do
Nothing urgent. A chief executive’s departure does not change a database contract or a product roadmap overnight, and the returning interim leader knows the company better than anyone. Customers should watch the Investor Day materials and the choice of permanent CEO for any shift in priorities, particularly on pricing for MongoDB Atlas and on its AI retrieval features.
What the market reaction says about key-person risk
MongoDB’s fall came on a day when the company reaffirmed its forecasts, so investors were pricing leadership, not trading results. A chief executive leaving after less than a year, with no permanent successor named, raises questions about strategy and about whether other senior staff will follow. Meta’s 4.8% fall the same day cannot be attributed to the announcement alone, since its shares move on many factors, but it is a reminder that investors are watching how much Meta spends on AI and when that spending will pay back.
Why Meta Enterprise Platform Matters to the AI Market
Meta has spent heavily on AI models, data centres and talent. CNBC and TechCrunch both note that investors want to see that spending pay off, and selling to businesses is one of the most direct routes to revenue.
Muse gives Meta a starting point
Meta Enterprise Platform launches with momentum. CNBC reports that Muse, released on 8 September, overtook ChatGPT as the leading free iOS app. Our coverage of the Muse agent’s dedicated VM tab and of Muse Code shows how quickly Meta has been shipping features. Consumer popularity does not guarantee enterprise trust, but it gives Meta’s sales team something employees already use.
Price is Meta’s sharpest weapon
Meta has competed on price since opening its model API. On OpenRouter, Muse Spark 1.3 costs $1.25 per million input tokens and $4.25 per million output tokens. A “contributor” tier costs $0.10 and $0.20 if the customer lets Meta use prompts and code to improve its models. The chart shows the gap.
The contributor tier is roughly 95% cheaper on output. For most enterprises it is also a non-starter, because it means sharing business data for training. We examined that trade-off when Meta launched contributor pricing for Muse Spark. Expect Meta Enterprise Platform contracts to be judged on the standard tier and on data terms.
Rivals already have a head start
OpenAI, Anthropic, Google and Microsoft all sell AI to enterprises through established sales teams, cloud partnerships and compliance programmes. Meta Enterprise Platform starts behind them on enterprise relationships and ahead of them on reach into small businesses through WhatsApp and Instagram. Hiring a chief executive who has spent his career selling to large enterprises suggests Meta intends to compete at the top of the market as well.
Security and Trust Questions for Meta Enterprise Platform
Desai says security and privacy are built in “from the outset”, citing Muse. Buyers will want evidence. Muse itself rolled out with a serious security flaw, as we reported on 23 September. Agents that read email, book travel and act for employees carry more risk than a chatbot, because a mistake is an action rather than a sentence.
Data use is the first question
Meta’s business is built on data. Enterprise customers will need contract terms that rule out using their data for advertising or model training, clear data residency options, and audit rights. Those terms are routine at established enterprise AI vendors, and their absence would slow Meta Enterprise Platform’s sales far more than any benchmark.
Questions Meta Has Not Yet Answered
The launch was an announcement of intent, not a product release. Several basics are still missing, and each will shape whether Meta Enterprise Platform is a serious option for large organisations.
Price and packaging
Meta has not said whether Meta Enterprise Platform will be sold as seats, as usage, or as a bundle with advertising spend. Its developer API and the WhatsApp agent are priced by tokens, which suits developers but makes budgeting harder for finance teams used to per-user licences.
Sales channels
Enterprise software is often bought through cloud marketplaces and resellers. Meta has not named partners, and it has not said whether Desai will build a direct sales force of the kind he ran at MongoDB and ServiceNow.
Where Muse fits at work
Muse is a personal agent built for consumers. Meta Enterprise Platform will need controls that let an employer decide what the agent may read, send and change, and who can see its activity. Until those administrative controls exist, Muse is better treated as a pilot than as a workplace tool.
What Businesses Should Ask Before Adopting Meta's Enterprise AI
Meta Enterprise Platform has no public price list or contract terms yet. Until it does, use these questions to compare it with the suppliers you already use. They belong in any vendor management review of a new AI supplier.
| Question | Why it matters |
|---|---|
| Is our data excluded from training and advertising? | The contributor tier shows Meta prices data use explicitly |
| Where is data stored and processed? | UK GDPR and EU GDPR obligations |
| What can agents do without human approval? | Agents that act carry more risk than chatbots |
| What service levels and support apply? | Consumer products rarely carry enterprise SLAs |
| How easily can we switch models or suppliers? | Avoid lock-in while the market moves monthly |
Start with a contained pilot
The lowest-risk route into Meta Enterprise Platform is a customer-messaging pilot with Meta Business Agent, where Meta already has a track record and a human hand-off. Keep general-purpose agents such as Muse out of sensitive systems until Meta publishes enterprise terms. If you are weighing Meta against other suppliers, an independent AI strategy review will compare total cost and risk rather than list prices alone.
Plan for a multi-vendor future
The most resilient approach is to build your AI agents and integrations so the underlying model can be swapped. Meta’s low prices make it worth testing, and the market’s pace makes it unwise to depend on any single supplier.
What Meta Enterprise Platform Means for UK Businesses
For UK organisations, the practical questions are about data. Any AI supplier handling customer conversations or employee email becomes a data processor under UK GDPR, so contracts need processing terms, a clear list of sub-processors and an answer on international transfers. Meta already provides processing terms for its business messaging products, and buyers should expect Meta Enterprise Platform contracts to follow the same pattern.
There is also a cybersecurity question. Agents that act on behalf of staff widen the attack surface, because a compromised or manipulated agent can send messages and change records. Ask how Meta Enterprise Platform logs agent actions, how quickly suspicious behaviour is flagged, and whether your own security team can see those logs.
What Happens Next
Desai starts immediately. The first signs of Meta Enterprise Platform’s direction will be enterprise contract terms, a price list, and named launch customers. MongoDB’s next milestone is its choice of permanent chief executive. For now, the clearest message from 28 September is that Meta intends to sell AI to businesses directly, and is willing to pay for experienced leadership to do it.
Meta Enterprise Platform FAQs
What is Meta Enterprise Platform?
It is Meta’s new business unit, announced on 28 September 2026, that will sell Meta’s AI stack to businesses and developers, starting with the Muse agent, Meta Business Agent, the Muse API and Muse Code.
Who leads Meta Enterprise Platform?
Chirantan “CJ” Desai, formerly president and CEO of MongoDB, as Meta’s Chief Enterprise Platform Officer. He reports directly to Mark Zuckerberg.
Who is running MongoDB now?
Dev Ittycheria, who led MongoDB from 2014 to 2025, is interim president and CEO while the board searches for a permanent successor.
How much did MongoDB’s shares fall?
They closed down 18.5% on 28 September 2026 at $334.68, after touching $300.00, 26.9% below the previous close. Meta’s shares closed 4.8% lower.
Can businesses buy Meta Enterprise Platform today?
Meta Business Agent, the Muse API and Muse Code are already available separately. Meta has not yet published enterprise pricing or contract terms for Meta Enterprise Platform as a whole.
References
Launching Meta Enterprise Platform (Meta Newsroom)
MongoDB Announces CEO Transition (MongoDB)
Meta launches enterprise AI platform, hires MongoDB CEO to lead new initiative (TechCrunch)
Meta hires MongoDB CEO CJ Desai for enterprise unit (CNBC)
Meta enterprise AI platform: MongoDB CEO CJ Desai departure (Quartz)
MongoDB, Inc. Announces Second Quarter Fiscal 2027 Financial Results (Nasdaq)
Meta’s AI agent for WhatsApp Business is now available globally (TechCrunch)
More AI coverage: explore Progressive Robot's AI Models, Tools & Releases hub — hands-on reviews, setup guides and benchmarks in one place.
