Prisma AIRS

Enterprise AI security platform from Palo Alto Networks for inspecting AI supply chains, testing applications and agents, and enforcing protection during live AI interactions.

About Prisma AIRS

Prisma AIRS is Palo Alto Networks’ platform for securing enterprise AI applications, models, agents, and infrastructure. Its capabilities span AI inventory, supply chain inspection, red teaming, runtime security, gateway controls, and agent identity. The platform supports security work before deployment and during ongoing operation.

AI Supply Chain Security, previously named AI Model Security, examines models and agent skills for risks before they enter production. AI Red Teaming uses information about a target system to generate relevant adversarial tests. Runtime Security inspects live interactions and gives teams tools for reviewing violations and applying policies.

Prisma AIRS documentation describes cloud, hybrid, and on-premises environments. Licensing uses Software NGFW credits, with consumption and deployment requirements depending on the selected capability.

Key Features

AI inventory: Maps AI workloads, resources, and traffic to improve visibility across enterprise environments.
AI supply chain scanning: Inspects models and agent skills for malicious content, unsafe components, and policy violations.
Contextual red teaming: Profiles target endpoints and generates attacks relevant to their configuration and tools.
Runtime threat protection: Inspects live AI interactions for prompt attacks, sensitive-data exposure, and other threats.
Investigation and policy controls: Provides session and violation views for reviewing events and adjusting protection.
Gateway and agent identity capabilities: Adds centralized AI traffic controls and identity governance for agent workflows.

Pricing

Licensing model: Bring-your-own-license deployment funded through Software NGFW credits. Public monetary rates were not established.
AI Red Teaming: Requires Software NGFW credits for access to automated testing.
Runtime API consumption: API-intercept usage is calculated through monthly token allocations, with the quota resetting at the end of each calendar month. The documentation defines a token as four characters.

Use Cases

Inspect externally sourced models and agent skills before deployment.
Test AI applications against attacks tailored to their tools and configuration.
Monitor production AI traffic and investigate policy violations.
Apply centralized access and security controls across enterprise AI services.

Pros & Cons

Pros:

Covers supply chain, testing, runtime, and governance requirements within one platform.
Supports contextual testing of applications and agentic endpoints.
Documents cloud, hybrid, and on-premises deployment scenarios.

Cons:

Deployment requires license activation and configuration of Software NGFW credit resources.
The separately promoted Managed AI Runtime Security offering is labeled preview or early access.

FAQ

What is Prisma AIRS?

Prisma AIRS is Palo Alto Networks' enterprise platform for securing AI applications, models, agents, and infrastructure.

What happened to AI Model Security?

The capability is now presented as AI Supply Chain Security, with coverage for models and agent skills.

How does Prisma AIRS red teaming work?

It profiles a target endpoint and generates adversarial tests informed by the system's configuration, tools, and intended use.

What does Runtime Security monitor?

It inspects live AI interactions, including prompts, responses, and associated data flows, and provides views for investigating violations.

Can Prisma AIRS support hybrid environments?

Yes. Its documentation describes support for cloud, hybrid, and on-premises environments.

Is Managed AI Runtime Security generally available?

The reviewed product page labels that managed offering as preview or early access. It should be distinguished from the broader Runtime Security capability.

Categories:

AI Security Software Security Software