Palo Alto Networks (NASDAQ: PANW), a global leader in cybersecurity, has strengthened its protection platform for AI environments with the arrival of Prisma® AIRS™ 3.0. This new version enhances security around agentic AI, enabling organizations to confidently advance toward using autonomous agents and securely manage their entire lifecycle. As a result, companies move from merely supervising AI interactions to safely authorizing their autonomous actions.
Adopting AI-driven business models presents significant security challenges, from uncontrolled shadow AI to critical areas like agent identity, runtime protection, and automated governance. While many companies monitor what AI “says,” they still lack visibility into what it actually does. Prisma AIRS 3.0 bridges this gap by providing comprehensive visibility and protecting agents from their design phase through deployment, even when they perform complex autonomous tasks.
Anand Oswal, EVP of AI and Network Security at Palo Alto Networks
“Agentic AI represents a major advancement, moving beyond simple conversations to autonomous action, which will redefine productivity. However, this shift from an AI that talks to an AI that acts introduces new risks — from unmanaged agent identities to unpredictable behavior at runtime. Prisma AIRS 3.0 offers an integrated platform to detect, assess, and safeguard agentic AI, empowering our customers to scale their AI-driven enterprises with confidence and security.”
Prisma AIRS replaces fragmented point solutions with a unified platform that manages the key threats and risks of AI applications and autonomous agents. Its new features prepare operational teams for the future as agent ecosystems evolve:
- Discover AI agents wherever they reside. Organizations can now generate an instant inventory of AI agents, models, and connections across their environment. Prisma AIRS identifies agents running in cloud environments, SaaS platforms, and locally on endpoints that traditional tools often overlook.
- Continuous risk assessment of AI agents. Security teams no longer need to guess whether an agent is safe. Agent Artifact Security analyzes an agent’s architecture to identify vulnerabilities. AI Red Teaming simulates agent-specific attacks considering context, detects AI-related vulnerabilities, and recommends runtime security policies.
- Real-time, scalable protection of AI ecosystems. The “AI Agent Gateway,” currently in limited preview, provides a centralized control plane to secure agent identity, runtime, governance, and observability. Once the proposed Koi acquisition is complete, “Agentic Endpoint Security” will offer the visibility needed to protect various AI applications on endpoints, allowing teams to deploy tools like encoding agents quickly and securely.
“To scale agentic AI, fundamental obstacles related to visibility and control must be addressed. Organizations need to know where their agents are, what data they access, and how they behave after deployment. By offering a comprehensive view of the AI lifecycle, Prisma AIRS helps clients mitigate risks. We are proud to partner with Palo Alto Networks to help businesses establish consistent security measures that enable safe expansion of agentic AI.”