Opsin, a provider of risk protection for enterprise AI agents, has closed a $21m funding round as companies face a widening gap between what autonomous agents can do and what existing security tools can monitor.
SYN Ventures and Race Capital led the investment. Opsin views the backing as evidence that the market accepts a fundamental change in how agent security must work.
The company argues that traditional enterprise defences were designed around endpoints, networks and the exchange of prompts and responses. Agents, by contrast, operate on servers, link straight into corporate data and hold authority to take action. Opsin contends that older tools cannot assess whether an agent’s conduct remains consistent with the task it was created to perform.
Opsin originally focused on helping businesses control excessive data sharing as AI adoption grew. As agents became more widespread, it extended its scope beyond safeguarding data in motion to overseeing the identities, intents, connectors, tool calls and permissions of the agents themselves.
To manage the inconsistent ways AI platforms categorise agents, Opsin built a standard model that normalises agents across enterprise-approved systems including Microsoft 365 Copilot, Copilot Studio, Claude Enterprise, ChatGPT Enterprise and Gemini. After identifying thousands of projects and recurring oversharing patterns, the team made agent posture, covering connectors, tool calls and sharing scope, and runtime behaviour central to how it builds context.
Examples of risky runtime behaviour include agents reaching beyond their role scope or delivering sensitive data to users whose roles do not match.
That context feeds the Opsin Risk Graph, which learns continuously from agent activity to detect new risks. When an agent breaches policy, Opsin can isolate it, stop the action or alert the agent’s owner about the required fix. The aim is to let security teams see every agent, understand why its behaviour poses a risk, decide which issues come first and hold evidence to support those decisions.
Looking three years ahead, Opsin expects agents to interact with one another far more than with humans, with much of that exchange never reviewed by a person in real time. The company believes the answer is not to restrict agents but to oversee them as organisations oversee employees, judging them on their actions and the demands of their role.
Opsin connects through an API without requiring specialist training and delivers an initial risk assessment of agents across approved AI platforms within 24 hours. Its customers range from mid-sized firms to Fortune 500 companies in healthcare, financial services, manufacturing and other regulated sectors.
The new capital will support faster product development, expansion among regulated enterprises, further growth of its risk protection layer as agents gain autonomy, and recruitment of additional talent.
Copyright © 2026 RegTech Analyst
Copyright © 2026 RegTech Analyst
