Opsin Raises $21M to Monitor Enterprise AI Agents

AI security and governance startup Opsin has raised $21 million to help organizations manage and monitor the growing use of autonomous AI agents in the workplace. The funding round was led by SYN Ventures and Race Capital, and the money will support expansion of Opsin’s sales and engineering teams as demand increases for tools that can track what AI agents are doing inside corporate systems. Opsin was founded by CEO James Pham, along with cofounders Jeremy Mailen and Oz Wasserman. The company focuses on a rapidly emerging problem: non-technical employees are increasingly deploying AI agents from providers such as OpenAI and Microsoft into business workflows, code and databases. These agents can complete tasks autonomously, but they also create visibility and security risks if they access sensitive information or take actions that users do not fully understand. Pham said the company was motivated after hearing repeated concerns that AI agents could misuse personal or confidential data. Opsin’s platform is designed to show where AI agents are active, which tools and data they access, what they were originally configured to do, and which agents may require review or cleanup. The company also provides continuous monitoring and alerts for risky behavior. The startup’s focus reflects broader cybersecurity anxieties as businesses adopt AI more quickly. Investors and industry watchers see AI agents as a new attack surface, with more agents creating more opportunities for exposure or exploitation. Cybersecurity stocks have risen this year, and major AI companies including Anthropic and OpenAI are reportedly hiring aggressively in security-related roles. SYN Ventures partner Glenn Chisholm, who backed the round, said the industry is entering a period where the goal is to safely extract value from AI rather than avoid it altogether. Opsin is positioning itself as a governance layer that could make enterprises more comfortable using products such as Microsoft Copilot and ChatGPT Enterprise, even as AI agents become more numerous and harder to supervise.