Offroad launches with $7M to automate identity security with AI agents
Offroad Inc. launched today with $7 million in funding to build what it calls an agentic identity security team, using artificial intelligence agents to investigate and remediate access risks across human users, machine identities and AI agents.
Founded in September 2025 and based in New York and Tel Aviv, Offroad is targeting a problem that has outgrown manual review. Identity security is no longer simply a question of who has access, the company argues, but a continuous operational challenge of why access exists, whether it remains justified, how it is used and what breaks if it changes. The context needed to answer those questions is scattered across identity providers, human resources systems, software-as-a-service applications, cloud platforms, security tools, tickets and logs.
Rather than surfacing another dashboard of findings, Offroad’s agents gather that context, investigate posture and runtime issues together and then resolve them, either by acting directly where it’s safe or by routing high-impact changes to the right people with the context needed to decide.
The company is pitching AI agents as the reason the old model no longer holds. Autonomous identities now operate at machine speed across many systems at once, breaking the behavioral baselines that runtime detection tools were built on. Service accounts, application programming interface keys and machine identities often carry broad, persistent access with weak ownership and long lifecycles that most organizations lack the processes to govern.
