Mumbai, India — Palo Alto Networks advances its AI security platform with the launch of Prisma® AIRS™ 3.0. As organizations transition toward a future defined by autonomous agents, Prisma AIRS 3.0 secures the entire Agentic AI lifecycle – enabling enterprises to move from simply observing AI interactions to safely authorizing autonomous execution.
The shift toward an AI-powered enterprise introduces systemic security challenges – ranging from unmanaged Shadow AI to the critical new frontiers of agentic identity, runtime security, and automated governance. While many enterprises monitor what AI says, they remain blind to what AI does. Prisma AIRS 3.0 closes this gap, providing visibility and securing agents from design to runtime as they execute complex tasks independently.
Anand Oswal, Executive Vice President of AI & Network Security, Palo Alto Networks
“Agentic AI represents a massive leap forward, moving beyond simple conversation to autonomous action that will redefine productivity. But this shift from “AI that talks” to “AI that acts” introduces new risks – from unmanaged agentic identities to unpredictable runtime behaviors. Prisma AIRS 3.0 provides a comprehensive platform to discover, assess and protect agentic AI, giving our customers the unique ability to confidently, and securely, scale the AI-powered enterprise.”
Prisma AIRS replaces fragmented point solutions with a single platform to manage the primary threats and risks of AI apps and autonomous agents. The new capabilities allow teams to future-proof their operations as agent ecosystems evolve:
- Discover AI Agents Wherever They Live. Organizations can now instantly inventory AI agents, models, and connections across their entire environment. Prisma AIRS identifies agents running in cloud environments, SaaS platforms and locally on endpoints that traditional tools miss.
- Assess AI Agent Risk Continuously. Security teams can stop guessing if an agent is safe. Agent Artifact Security maps out an agent’s architecture and scans for vulnerabilities. AI Red Teaming for agents simulates context-aware agentic attacks, discovers AI-related vulnerabilities, and recommends runtime security policies.
- Protect AI Ecosystems in Real-Time at Scale. The AI Agent Gateway, currently available in limited preview, provides a central control plane to enforce agent runtime and identity security, governance and observability. Following the close of the proposed Koi acquisition, Agentic Endpoint Security will offer the visibility needed to secure various AI endpoint applications, empowering teams to deploy tools like coding agents at speed without sacrificing safety.
