Oasis Security Raises $120M for AI Access Mgmt

🔑 Enhanced Key Takeaways

• •Oasis launched Agentic Access Management (AAM™) on November 19, 2025, as the first identity solution purpose-built for governing AI agents across their lifecycle with intent inference, deterministic policy enforcement, and just-in-time session identities.[2]
• •AAM integrates with platforms like Cursor via partnerships, enabling intent-based access, policy enforcement through Cursor Hooks, and a centralized Shadow AI Dashboard for enterprise visibility into supply chain risks and asset exposure.[5]
• •The platform supports securing AI agents across diverse environments including AWS, Azure, GitHub, ChatGPT, Salesforce, and Office 365, extending from non-human identities to top-layer AI agents.[8]

🛠️ Technical Deep Dive

• •Intent Inference: Uses LLM-driven analysis to interpret agent prompts, identifying intent, invoker, data access, and purpose for policy evaluation.[2]
• •Deterministic Policy Enforcement: Policy engine validates requests in real-time, escalates with human-in-the-loop for high-privilege actions, ensuring repeatable outcomes.[2]
• •Just-in-Time (JIT) Session Identities: Provisions ephemeral, least-privilege credentials for short sessions (seconds to minutes), with automatic teardown and full audit trails linking human, prompt, policy, actions, and results.[1][2]
• •Chain of Custody: Tracks full traceability from Human → Agent → Prompt → Intent → Policy → Identity → Actions → Results for every session.[1]
• •Context-Aware PAM Escalation: Applies time-bound, policy-driven privilege elevation based on business context and risk level.[3]

🔮 Future ImplicationsAI analysis grounded in cited sources

⏳ Timeline