AI tools can spread through a company faster than its security team can assess them. Pluto Security says a newly announced tool may appear inside an organization within two hours, while apps built by employees can expose sensitive information or give agents more access than intended.
The startup has raised $42 million to help businesses see and manage those risks. Its approach follows a shift from tracking what employees do toward watching the actions of the AI agents and tools they use.
Employee-built apps create fresh exposure
Pluto CEO and co-founder Shahar Bahat told TechCrunch that about 60% of the risks the company detects across its customers involve tools built by non-developers using AI. Finance teams are the largest source of those risks, followed by marketing and legal. Developers rank fourth.
One example involved a finance employee at a Fortune 500 company who built an app to manage customer information and churn. The app was left exposed to the internet. The case shows how an employee can create a useful tool quickly while also introducing a security problem the company may not have anticipated.
The risk extends beyond the apps themselves. AI tools may rely on skills, extensions, connectors and other components, and agents can take actions as they work. Companies need visibility into those pieces and their behavior, Pluto argues, rather than treating AI use as a simple matter of tracking which software is installed.
Monitoring what agents do
Pluto’s software monitors the AI tools employees use, the components those tools rely on, and the apps and agents people create. The company says it can identify risky components, flag potentially dangerous agent activity and enforce security policies while allowing employees to keep using or building AI tools.
To gather information, Pluto connects to endpoint security and device management systems. It can deploy sensors through those systems or directly onto devices. The information it collects includes prompts, commands, running processes, permissions and users’ previous activity. Lightweight models running on endpoints help assess what an agent is trying to do.
The system can also intervene during an agent’s work. If an agent attempts to obtain permissions beyond what the user intended, Pluto can stop the action or add instructions to the session that ask the agent to seek confirmation. Bahat described this as a way to change how an agent is working by injecting context into its session.
That capability reflects the company’s central premise: security teams must pay attention to agent behavior as well as user behavior. The challenge is to detect and constrain risky actions without blocking all AI use across an organization.
Early traction and a public research platform
Pluto began selling its product in April and says it has passed $2 million in annual recurring revenue. It reports more than 30 enterprise customers, including Oscar Health, Workato, Sixt and AlphaSense.
The company’s funding includes an $8.5 million seed round led by TLV Partners, with investment from the founders of cybersecurity firm Armis. Its $34 million Series A was led by Jump Capital, with Skinos, Ansa Capital and other angel investors participating.
Pluto also launched Plutonium in April, a free platform for checking risks in publicly available AI skills, connectors and add-ons. It includes an MCP scanner and a catalog of verified components. Bahat said thousands of security professionals use the platform and that its community provides product feedback and prospective customers. She said 70% of the company’s sales pipeline now comes from inbound interest, without specifying how much is directly tied to Plutonium.
Where Pluto plans to grow
The startup plans to use its new funding to develop its product, expand security research and grow its presence in the U.S. It has 40 employees in the U.S. and Israel.
For companies adopting AI, Pluto’s pitch is that oversight needs to cover the tools employees select, the components those tools depend on and the actions agents take. As AI use spreads across departments, security teams face the task of finding hidden exposure while preserving room for employees to work with the technology.