As artificial intelligence tools become ubiquitous in the workplace, enterprises are facing a new frontier of security vulnerabilities. The rapid, often unmonitored adoption of AI applications by non-technical staff has created significant blind spots, with departments like finance, marketing, and legal frequently deploying AI-driven tools without traditional IT oversight. Pluto Security, a startup emerging from stealth, is positioning itself to address these gaps by providing visibility and control over how AI agents and extensions operate within corporate environments.

Led by CEO and co-founder Shahar Bahat, Pluto Security has already demonstrated significant market traction, reaching $2 million in annual recurring revenue shortly after launching its sales operations in April. The company’s platform monitors AI tools, extensions, and agent behavior in real-time, allowing security teams to identify risky components and enforce policies without stifling employee productivity. By integrating with existing endpoint security systems, Pluto can intervene during active sessions, preventing agents from overstepping their permissions or accessing sensitive data.

The startup’s approach is bolstered by a successful $42 million funding haul, comprising an $8.5 million seed round and a $34 million Series A. Investors include TLV Partners, Jump Capital, and notable cybersecurity figures such as the founders of Armis and Check Point co-founder Shlomo Kramer. This capital will be used to scale operations in the U.S. and Israel, expand security research, and further develop the platform’s ability to analyze agent behavior.

Beyond its core enterprise offering, Pluto has gained industry attention through ‘Plutonium,’ a free platform that allows security professionals to scan for risks in publicly available AI skills and connectors. This community-driven tool has become a vital lead-generation engine, with a significant portion of the company’s sales pipeline originating from inbound interest. As organizations move from monitoring user behavior to managing autonomous agent activity, Pluto’s real-time intervention capabilities represent a shift in how companies approach the security of their AI-integrated workflows.

Key Takeaways

• Pluto Security raised $42 million to help enterprises manage the security risks associated with the rapid, decentralized adoption of AI tools.
• The platform focuses on monitoring 'agent behavior' in real-time, allowing it to intervene and block unauthorized actions or permission escalations during active sessions.
• The company has already achieved $2 million in annual recurring revenue, with a client base that includes major firms like Oscar Health and Workato.

Editor’s Analysis & Impact

The rise of Pluto Security highlights a critical transition in the cybersecurity landscape: the shift from securing static software to governing dynamic, autonomous AI agents. As non-technical employees increasingly build and deploy their own AI-powered workflows, traditional perimeter-based security is becoming obsolete. The market is clearly signaling a demand for ‘AI-native’ security tools that can interpret context and intent rather than just signature-based threats. Pluto’s strategy of combining a high-end enterprise product with a free, community-facing tool (Plutonium) is a savvy move to establish industry authority and build a robust sales pipeline. Looking ahead, the ability to provide ‘guardrails’ for AI agents without hindering innovation will be the primary differentiator for security vendors. Companies that fail to implement such visibility will likely face significant data leakage and compliance risks as AI adoption continues to outpace internal governance.

Frequently Asked Questions

Q: What is the primary security risk Pluto Security addresses?
A: Pluto Security addresses the risks posed by employees using AI tools, extensions, and agents without IT oversight, which can lead to exposed sensitive data and unauthorized system access.

Q: How does Pluto Security intervene in AI agent activity?
A: The platform monitors AI tools in real-time and can inject context into a session to stop an agent from performing unauthorized actions or prompt the agent to request user confirmation for sensitive tasks.