Rig Security, a Tel Aviv identity security startup, came out of stealth Tuesday with $12 million in seed funding to help companies see and control what AI agents do under employees' credentials, SecurityWeek reported. The round was led by Ten Eleven Ventures and Brightmind Partners.

What Happened

CrowdStrike's Falcon Fund and a group of angel investors also joined the round. Among the angels is Ami Luttwak, co-founder and CTO of cloud security company Wiz. Rig was founded in 2025 by chief executive Guy Kozliner, who previously worked on Wiz's CTO team.

The company's platform rests on two components. RICE, the Rig Identity Correlation Engine, uses machine learning and mathematical modeling to resolve which accounts across different systems belong to the same identity and to build a graph of dependencies between them. Rig claims it does this with more than 96% accuracy. Bifrost is a lightweight endpoint sensor that distinguishes AI agent sessions from genuine user activity and enforces policy on them.

RelatedProofpoint to Buy Varonis for $6.3 Billion, Citing Risks of AI Agents' Data Access

Around those two pieces, Rig offers identity security posture management, identity threat detection and response, AI discovery and attribution, runtime policy enforcement and non-human identity security. The company says the product is already deployed at Fortune 200 organizations in financial services, insurance, healthcare and technology.

Why It Matters

Rig is built around a specific blind spot. When an employee launches an AI agent, the agent typically inherits that person's identity and permissions, or those of the service that started it, and then operates at machine speed on its own. Logs show the account took an action, but not whether a person or the agent did it. Kozliner told SecurityWeek that security teams also cannot stop the agent without blocking the human, because disabling the account cuts off both.

That makes common response playbooks awkward. Suspending a compromised account works when the account belongs to one actor. When a legitimate employee and an unmonitored agent share it, the same step disrupts real work, while leaving the account active leaves the agent unchecked. Kozliner described the underlying risk as access that is broader than intended and ungoverned, "hiding behind human activity."

Separating the two at the endpoint, as Bifrost aims to do, would let a security team apply a policy to agent sessions alone, for example blocking a risky action by the agent while the employee keeps working.

A Crowded Category

Agent identity has quickly become one of the busiest areas in security. Established identity vendors and cloud platforms have been adding controls for non-human and agent identities, and several startups have raised money on similar pitches in the past year. Rig's angle is correlation: tying scattered accounts back to a single identity and then attributing activity within that identity to either a person or an agent.

The investor list signals ties to two large security platforms. CrowdStrike's venture arm is on the cap table, and Rig's founder and one of its angels come from Wiz, which Google agreed to acquire in 2025.

What to Watch

Rig said it will use the money for research and engineering, especially detection and enforcement, for building a U.S. sales team, and for channel partnerships. The 96% correlation accuracy figure is a company claim, and customers will want to see how Bifrost performs as agents run from more places than a managed endpoint, including cloud-hosted agent environments.