Pluto Security secures $42M for enterprise AI agent oversight
Pluto Security has emerged from stealth with $42 million in funding for software designed to help enterprises observe and control employee use of AI tools and AI agents. The startup says it began selling in April, has passed $2 million in annual recurring revenue, and now serves more than 30 enterprise customers, including Oscar Health, Workato, Sixt and AlphaSense.
The financing includes an $8.5 million seed round led by TLV Partners, with investment from the founders of cybersecurity company Armis. Pluto also raised a $34 million Series A led by Jump Capital, with Skinos, Ansa Capital and angel investors participating.
AI adoption expands the enterprise attack surface
CEO and co-founder Shahar Bahat said a newly announced AI tool can enter a company environment within two hours of its release. Pluto’s experience across its customer base suggests that the exposure is not confined to software engineering teams: the company says about 60% of the risks it detects arise from tools vibe-coded by non-developers.
Finance departments are the largest source of those risks, Bahat said, followed by marketing and legal teams, while developers rank fourth. In one case, Pluto found an application built by a finance employee at a Fortune 500 company to manage customer information and churn; the application had been left exposed to the internet.
The company says it has found malicious items in every proof of concept it entered, including extension packages and skills. Some had active live attacks while the proof of concept was underway. The pattern underlines why AI adoption can create security work outside conventional development and procurement processes.
From user activity to agent activity
Pluto’s approach is based on monitoring what agents do in real time as well as tracking the AI tools, applications and agents employees build and deploy. Its platform covers AI skills, extensions, MCP servers and other components used by those tools, aiming to identify risky components, flag potentially dangerous activity and enforce policies without preventing employees from using or creating AI tools.
To collect signals, Pluto connects to endpoint security and device-management systems and can deploy its own sensors through those systems or directly to devices. It gathers information such as prompts, commands, running processes, permissions and prior user activity. The company also uses lightweight models on endpoints to assess what an agent is attempting to do.
Pluto says its system can intervene during an agent session. If an agent seeks permissions beyond a user’s intended scope, the platform can stop the action or add instructions that require the agent to ask for confirmation. This makes the control point more immediate than a record of activity reviewed after the action is complete.
Product development and market reach
The startup also operates Plutonium, a free platform launched in April for checking risks associated with public AI skills, connectors and add-ons. It includes an MCP scanner and a catalogue of verified components. Bahat said thousands of security professionals use the platform, and Pluto says 70% of its sales pipeline now comes from inbound interest.
The funding will support product development, security research and expansion in the United States. Pluto has 40 employees across the US and Israel. Its position in the market sits alongside enterprise AI agent deployment models as enterprises assess how to deploy AI agents while retaining operational oversight and control.
For businesses, Pluto’s model points to a practical requirement: AI governance needs to cover employee-created applications, third-party components and live agent permissions, not solely approved tools or traditional developer workflows.