Mark Zuckerberg built one of the most profitable advertising businesses in history. He is now building a second business from scratch.
Meta is planning to form a new unit called Enterprise Solutions, designed to place engineers and product managers directly inside large corporate customers to help them deploy Meta’s artificial intelligence (AI) tools.
According to a report by The Information, senior executive Naomi Gleit detailed the structure in an internal memo: product managers to lead client engagements, data engineers to prepare corporate data for Meta’s AI systems and software engineers to wire Meta’s products into the systems companies already run.
The enterprise technology playbook is familiar. The logo on it is not.
The Playbook Meta Is Borrowing
The approach copies what is now a standard move across the AI industry. Google Cloud CEO Thomas Kurian announced this month that the company would build a team of forward-deployed engineers embedded with enterprise clients. OpenAI launched a majority-owned subsidiary called OpenAI Deployment Company, backed by more than $4 billion in initial investment, to put engineers on-site at corporations. Anthropic formed a parallel venture with Wall Street financial firms in early May, structured around the same embedded-engineer logic. Salesforce has publicly committed to 1,000 forward-deployed engineers of its own.
The model can be traced back to Palantir, which built its enterprise business over two decades by sending engineers into client organizations rather than selling software through a catalogue. Every major AI lab now treats that approach as table stakes.
Advertisement: Scroll to Continue
Meta joins late. The gap between the company and its competitors in enterprise AI isn’t primarily technical; it’s institutional. Meta has tried to sell to corporations before and has little to show for it.
A Record Built on Misses
In 2016, Meta launched Workplace, a business-focused version of Facebook intended to compete with Slack and Microsoft Teams. It shut the product down earlier this year after nearly a decade of declining relevance, citing the need to focus on AI. Workplace struggled with data security concerns and failed to integrate with the enterprise tools that companies actually used. Meta also runs a business messaging operation on WhatsApp and Messenger, though majority of the company’s revenue still comes from advertising.
That dependency is increasingly visible on the balance sheet. Meta raised its full-year capital expenditure guidance to between $125 billion and $145 billion earlier this quarter, up from an earlier range of $115 billion to $135 billion.
What Meta Is Betting On
The Enterprise Solutions unit is one answer to that pressure. Gleit wrote in her memo that the team will build “repeatable playbooks and tooling” so the work can scale. It is the standard language of a company trying to productize a service model it hasn’t yet tested.
The unit’s formation comes alongside a separate internal reorganization. According to The Information, Chief Technology Officer Andrew Bosworth outlined in a memo last month an initiative called the Agent Transformation Accelerator, which shifts employees from doing tasks to overseeing AI agents that carry them out. Bosworth called 2026 “a critical year” for Meta’s transformation.
Meta also announced that its testing paid subscriptions for its AI chatbot across Facebook, Instagram and WhatsApp, adding another potential revenue stream alongside the enterprise push. The company simultaneously announced layoffs affecting roughly 8,000 workers and moved more than 7,000 employees into new initiatives, with Enterprise Solutions among them.
