Opsin raises $21M to govern AI agents across enterprise platforms

Opsin closes $21 million to contain agent risk

Opsin, an AI governance and security platform, said October 15, 2026 that it raised $21 million in funding led by SYN Ventures and Race Capital. The San Jose company said the round will support enterprise adoption of goal oriented AI agents, which can act across systems to complete assigned tasks and may take extreme or unauthorized actions when permissions are broad and human oversight is limited.

From permissions to intent

Opsin said its platform connects an agent's purpose, permissions, data access, tools and activity across Microsoft Copilot, Claude, ChatGPT Enterprise, Gemini and other enterprise approved AI tools. The company said this gives security teams context to spot access or behavior outside an agent's role, prioritize significant risks and address issues at the source before they become incidents.

In an analysis of enterprise AI agent activity, Opsin found 93 percent of misaligned tools connected to agents had full, unrestricted access, and only 1 in 200 required human approval before acting.

Founders and customers

Opsin was founded by MIT alumnus James Pham, cybersecurity veteran Oz Wasserman and engineering leader Jeremy Mailen. Pham and Wasserman previously worked together at Abnormal Security. The company said it already serves nearly 20 enterprise customers, including Culligan, Barry-Wehmiller, Wellstar Health System and UiPath, across healthcare, financial services, manufacturing and other regulated sectors.

Investors frame the trust gap

Race Capital general partner Alfred Chuang said agents are gaining access and authority to act on critical systems, and Opsin links what an agent is supposed to do with what it actually does. SYN Ventures partner Glenn Chisholm said the platform is intended to give enterprises confidence to use AI agents without opening the door to uncontained risks. With the new funding, Opsin plans to expand the platform as employees increasingly supervise teams of agents working across enterprise data, knowledge and systems.

Sources: Newswire · Techmeme