TL;DR
- Enterprise Unit: Meta is said to be planning Enterprise Solutions to place engineers and product managers inside large corporate customers.
- Rival Model: OpenAI and Anthropic already pair AI rollouts with deployment teams, while Microsoft emphasizes governance controls for enterprise agents.
- Buyer Unknowns: Meta has not publicly confirmed customers, staffing scale, pricing, or rollout timing for the reported plan.
Meta is reportedly planning a new enterprise unit that would place engineers with large customers and widen business use of its AI tools. Product managers would join the same accounts through the same unit. Customer names, staffing scale, pricing, and rollout timing remain undisclosed.
Large companies often hit the same wall after an AI demo works: integration work, security review, approval chains, employee training, and workflow redesign can still stop a broader rollout. Embedded vendor staff would push Meta beyond selling model access toward helping customers clear those operational barriers. More direct involvement makes the plan more consequential than a routine packaging change.
How Meta Would Sell AI Into Large Businesses
A memo attributed to Naomi Gleit, Meta’s head of product, names the reported unit Enterprise Solutions.
“The new organization, called Enterprise Solutions, will place engineers and product managers inside large corporate customers.”
Naomi Gleit, Meta’s head of product (via The Information)
Enterprise Solutions will put Meta staff inside the technical and product decisions that often determine whether an AI rollout stays small or becomes part of daily operations. Engineers can address reliability and integration problems early, while product managers can turn customer requests into packaging changes or roadmap priorities. For large buyers, those steps can matter more than model quality alone because compliance reviews, data restrictions, and internal approvals often slow adoption after a promising pilot.
Working inside customer accounts would also shift execution risk back toward Meta once a buyer moves past the evaluation phase. Teams operating that close to a rollout could surface permission conflicts, workflow failures, and feature gaps before they harden into stalled deployments, while product managers could decide which requests become supported features, service terms, or pricing changes. Procurement and IT groups would get a clearer line of accountability, and Meta would gain a direct role in the commercial and operational decisions that often determine whether a pilot becomes a broader contract.
Rivals Already Sell Enterprise AI With Services or Controls
Meta follows the same strategy as its competitors. OpenAI on May 11 launched the OpenAI Deployment Company after an enterprise consulting push built around embedded specialists. Forward Deployed Engineers in that model work inside customer organizations on operational problems that can block a broader rollout.
Anthropic moved in a similar direction on May 4, when it formed a new AI services company with financial partners to support long-term Claude deployments. Microsoft is addressing the same enterprise bottleneck through governance rather than staffing alone, with Agent 365 managing agent inventory and permissions across enterprise environments. Together, those approaches show how major AI vendors are trying to stay closer to customer operations after the initial model sale, whether through embedded teams or tighter administrative control.
Rival models also show that enterprise AI spending is shifting beyond headline model access. OpenAI and Anthropic are putting implementation capacity closer to the customer, while Microsoft is emphasizing the controls that large organizations need before they let more agents run across sensitive systems. Meta’s reported unit would fit that same market logic even if its exact service design is still unclear.
Meta’s Earlier AI Buildout Leaves Open Questions
Earlier in 2026, Meta’s broader AI adoption drive made AI use part of employee reviews before this reported enterprise unit surfaced. Meta also created an infrastructure organization to expand AI computing capacity. Both moves show the company was already building dedicated AI structures inside its own operations while preparing for heavier deployment demands.
Earlier changes do not confirm how large Enterprise Solutions would be, but they make the reported direction easier to read. A company pushing employees to use AI more deeply and investing in more computing capacity would also have reason to build a higher-touch team for large business customers. Internal reorganization alone does not create a services business, yet it does show that Meta was already moving resources toward more demanding AI use cases.
Major details still separate the reported plan from a concrete buying option. Meta has not named customers, staffing commitments, service packaging, or rollout timing, and public detail is still missing on how deeply those teams would work inside a client’s technical or product operations. A public reference customer or formal staffing commitment would be the clearest sign that Enterprise Solutions is becoming a real services business rather than a limited support layer.
