The news
Omnissa, the former VMware end-user computing business, announced a set of Omnissa AI agents for IT teams on September 29, along with a managed Omnissa Cloud PC service and Omnissa Elara, a product for governing how employees and software agents use AI.
Three of the new agents target routine IT work, according to the company. A Digital Employee Experience (DEX) agent investigates problems employees run into, traces the root cause and recommends fixes drawn from playbooks IT already uses. A Workspace ONE Vulnerability Defense Agent reviews and ranks security vulnerabilities and prepares patching actions, with a person kept in the loop. A Windows App Lifecycle Agent handles application packaging and testing, then presents the results to an administrator.
Omnissa also introduced Horizon Delegate, an agent that carries out tasks for users of its Horizon virtual desktops under their approved identity and permissions, and that keeps working after the user goes offline. A hosted Omnissa MCP Server lets administrators connect outside AI agents to Omnissa services, limited to what their existing roles allow. MCP, or Model Context Protocol, is a common standard for linking AI agents to software tools.
Omnissa Cloud PC packages a cloud-hosted desktop with the company's endpoint management, digital employee experience tools and Horizon Cloud desktop-as-a-service offering. It will run first on Amazon Web Services, with other cloud providers to follow. Omnissa did not give availability dates for the new products, SiliconANGLE reported, and the company's announcement did not include pricing.
Elara is the governance piece. Omnissa says it gives IT visibility into which AI apps, models and agents are in use and lets teams apply policies based on business conditions; SiliconANGLE described it as applying those policies before high-impact actions run. Omnissa tied the launch to its State of the Digital Workspace report, which it said found AI assistant usage across enterprise endpoints rose nearly 1,000% year over year, with three-quarters of it coming from unsanctioned tools.
The products were announced at the Omnissa ONE 2026 conference in Orlando, according to SiliconANGLE. Omnissa was spun out of VMware in 2024 after Broadcom (AVGO) acquired VMware and sold its end-user computing business to KKR (KKR) for $4 billion, the outlet reported. Omnissa says it has 26,000 customers.
The numbers
- AI assistant usage on enterprise endpoints, year over year (Omnissa report)
- Up nearly 1,000%
- Share of that AI usage from unsanctioned tools (Omnissa)
- Three-quarters
- Endpoint data processed per day (Omnissa)
- About 50 TB
- Workflow executions per month (Omnissa)
- 150 million
- Omnissa customers (company figure)
- 26,000
- KKR purchase of VMware's end-user computing unit (2024, per SiliconANGLE)
- $4 billion
Why CEOs should care
For CIOs, the agents go after work that desktop support teams and outsourced service desks usually handle by hand: working out why an employee's laptop is misbehaving, preparing patches and packaging apps. Omnissa has published no prices or dates, so treat the launch as a roadmap rather than a product you can budget for. Useful questions for an Omnissa account team: which agents are generally available and when, how they will be priced (per device, per user or by usage), which actions each agent can take without human sign-off, and how every action is logged. If you are renewing a desktop support or managed services contract, ask the provider how its pricing would change if tools like these take over part of the ticket load.
CISOs should look hardest at Horizon Delegate, because it acts under an employee's identity and keeps going after that person goes offline. Ask how delegated tasks are scoped, how quickly they can be revoked, and whether they show up in audit trails as the user or as the agent. Omnissa's own framing raises the stakes: it says the time to exploit vulnerabilities has fallen to less than a day. That is the argument for an agent that prepares patches, but it is also a reason to check that the human review step does not become the new bottleneck.
Boards and CFOs should note the governance claim. If Omnissa's figure that three-quarters of AI assistant use on work devices comes from unsanctioned tools resembles your own company, company data may be flowing into tools nobody approved. Before buying Elara or any rival, ask IT for a count of AI tools in use today and which of them handle customer or employee data.
The bigger picture
Omnissa's pitch is that the platform that already manages employees' devices should also run and govern the agents working on them. It leans on data it already collects: the company says its platform processes about 50 TB of endpoint data a day and supports 150 million workflow executions a month. In a statement included in Omnissa's announcement, Phil Hochmuth, a research vice president at IDC, said autonomously managing and securing the end-user workspace has become a requirement for enterprise IT. Since the 2024 spinout, Omnissa has competed as a standalone company rather than as a unit of VMware.
What’s next
Watch for availability dates and pricing for the agents, Elara and Omnissa Cloud PC, and for which cloud providers follow Amazon Web Services. The more telling test will be whether early customers report fewer support tickets or faster patching once the agents are running, and how Elara's policies hold up against the AI tools employees already use without approval.
What “Fact-checked” means
Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.
- What we checked
- Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
- How
- A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
- Who
- The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, Hussein Mukhtar. A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
- If something is wrong
- “Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error
