Opsin Raises $21M to Keep AI Agents From Going Rogue

SAN JOSE, Calif. — Opsin, the AI governance and security platform built to detect and contain AI agent misbehavior, announced on October 15 that it has raised $21 million in funding to help enterprises securely adopt AI agents at scale, according to the company's press release carried by ACCESS Newswire. The round was led by SYN Ventures and Race Capital, and Opsin was founded by a team of former Abnormal Security leaders.

The announcement arrives as companies increasingly deploy goal-oriented AI agents — systems designed to independently take a series of actions to complete an assigned task, rather than simply respond to individual prompts. According to Opsin's announcement, incidents of agents going rogue are becoming more common as these systems gain broader access and autonomy inside enterprise environments. Business Insider, which identified the startup's cofounders as Jeremy Mailen, James Pham, and Oz Wasserman, reported that Opsin is positioning itself to complement the agent boom driven by OpenAI, Microsoft, Anthropic, and other platform providers.

Visibility and Control for Autonomous Agents

Opsin gives companies a way to keep tabs on how their agents are operating and to catch quickly when behaviors deviate from their intended purpose. The platform gives security teams centralized visibility and control over agents employees create across enterprise platforms, including Microsoft Copilot, Claude, ChatGPT Enterprise, Gemini, and other approved enterprise AI products.

The company connects information about an agent's purpose, identity, permissions, data access, tools, and activity to establish context around what the agent is supposed to do and how it actually behaves. That contextual approach helps security teams identify excessive access, discover AI agents operating across their organizations, detect behavior outside an agent's intended purpose, prioritize the most significant risks, and enforce security controls before problematic activity becomes a larger incident. According to coverage of the raise by Pulse2, Opsin plans to use the new capital to expand the platform as enterprises deploy AI agents that can independently access data, use software tools, and take action across business systems.

Agents that pursue goals can take extreme or unauthorized actions to achieve them — creating serious security risks, exposing sensitive data, bypassing company controls, or disrupting critical systems — particularly when they are given broad permissions with limited human oversight. Opsin's pitch is that visibility has to keep up with autonomy, and that security teams need a live picture of what each agent can do, not just what it was asked to do.

The Numbers Behind the Rogue-Agent Problem

In an analysis of AI agent activity across its enterprise customers, Opsin found that 93% of misaligned tools connected to agents had full, unrestricted access to the systems they were connected to. Just 1 in 200 required human approval before taking action — a figure the company says highlights how much freedom autonomous agents are being granted without human intervention.

Independent research points in the same direction. According to new research from OutSystems released with CIO Dive and KPMG, 82% of executives believe their organizations can keep pace with governing, integrating, and monitoring AI agents as they move into production — yet only 26% say they are well prepared to resolve conflicts across agents, systems, and automated workflows. The report frames that confidence gap as the defining challenge of the agentic era, arguing that safety comes from building control into the platform rather than reviewing it afterward.

The security exposure is not just theoretical. Cogent Research found that attack paths viable only for AI agents run at least twice as deep as paths within reach of human attackers, and that the average enterprise gained 34 new such paths in August 2026 — up 386% from a year earlier — according to reporting by Computer Weekly's Developer Network. Attackers' incentives have changed because agents don't tire: for a swarm of machine attackers, one more hop through an enterprise network costs almost nothing.

Agent Governance Becomes a Venture-Backable Category

Opsin's raise adds to a fast-growing wave of investment in agent infrastructure and oversight. This week also saw Workday tell investors it wants customers to fund its AI agents from staffing and outsourcing budgets rather than just IT spending, as reported by The Register — a sign that enterprises are already recalculating what agent labor is worth, and that the governance layer around it is becoming part of the deal.

The Opsin team brings credibility from the email-security world: the founders previously built at Abnormal Security, an AI-native security company, and the startup is combining that experience with a platform purpose-built for the agent era. SYN Ventures, one of the round's co-leads, specializes in cybersecurity investing — a fitting backer for a company arguing that the next big security perimeter is not the network edge but the permissions granted to software that acts on its own.

For enterprises, the message is blunt: every goal-oriented agent is an insider with ambitions, and most of them are operating with far more access than their jobs require. Opsin is betting that closing that gap — with continuous visibility, context-aware risk detection, and controls that trigger before incidents escalate — is now a standalone category worth funding. With $21 million in new capital and a leadership team that has fought the last generation of AI-driven attacks, the startup is moving early in what looks like a long race to govern the agents enterprises are already deploying.