At RSAC 2025, Palo Alto Networks launched its new Prisma AIRS (AI Security), a comprehensive security platform targeting threats across enterprise AI ecosystems. Building upon their “Secure AI by Design” portfolio introduced last year, Prisma AIRS addresses emerging security challenges posed by the proliferation of AI applications, agents, and models.
Core Capabilities
Prisma AIRS is an integrated security platform specifically targeting the unique threats posed by enterprise AI deployments across applications, agents, models, and data.
The platform provides a unified solution rather than requiring organizations to integrate several point products, reducing operational complexity while providing comprehensive protection across the AI lifecycle.
The platform includes five primary technical capabilities: AI Model Scanning, Posture Management, AI Red Teaming, Runtime Security, and AI Agent Security.
AI Model Scanning
This provides vulnerability detection for AI models, focusing on:
- Detection of model tampering evidence
- Identification of embedded malicious scripts within models
- Prevention of deserialization attacks
- Model integrity verification
Posture Management
This capability provides visibility into AI ecosystem security configurations, including:
- Permission auditing to identify excessive access rights
- Sensitive data exposure detection
- Platform and access misconfiguration identification
- Continuous compliance monitoring for AI systems
AI Red Teaming
Proactive security testing specifically adapted for AI systems:
- Automated penetration testing for AI applications and models
- Adaptive testing methodologies that simulate evolving attack patterns
- Stress testing of AI deployments under adversarial conditions
- Risk exposure measurement before production deployment
Runtime Security
Active protection during AI system operation against:
- Prompt injection attacks targeting LLM functionality
- Malicious code execution prevention
- Toxic content filtering
- Sensitive data leak prevention
- Resource consumption monitoring to prevent denial of service
- Hallucination detection and mitigation
AI Agent Security
Addresses emerging threats against autonomous AI agents:
- Identity impersonation prevention
- Memory manipulation detection
- Tool misuse prevention
- Security for agents developed on low-code/no-code platforms
Impact to IT Organizations
Prisma AIRS delivers real-world benefits to IT organizations:
- Consolidation of AI security functions, reducing the need to implement and maintain multiple point solutions
- Unified visibility across the AI security landscape, simplifying security operations
- Automated testing capabilities to reduce manual security assessment workloads
- Runtime protection helps prevent AI-specific incidents that traditional security tools might miss
- Compatibility with no-code/low-code platforms addresses the security challenges of democratized AI development
The solution will help IT teams across several scenarios:
- Enterprise AI Governance: Centralized visibility and control over all AI deployments
- Secure LLM Deployment: Model scanning prior to deployment, runtime protection during operation
- Autonomous Agent Security: Implementation: Continuous monitoring of agent activities and permissions
- AI Development Security: Integration with CI/CD pipelines for continuous model security testing
- Regulatory Compliance: Automated assessment of AI systems against emerging regulations
Analysis
The platform’s comprehensive approach addresses legitimate security challenges created by the rapid proliferation of AI applications, agents, and models across enterprises. The timing nicely aligns with the growing recognition of AI-specific security risks that traditional security tools inadequately address.
Palo Alto Networks is positioning Prisma AIRS as a comprehensive, integrated platform for securing enterprise AI ecosystems. This approach differentiates from competitors that offer point solutions addressing specific aspects of AI security.
Prisma AIRS builds on the company’s established credibility in enterprise security while extending into the emerging AI security domain. This gives Palo Alto Networks several key differentiators:
- End-to-end coverage across the AI lifecycle
- Integration of proactive (scanning, posture management) and reactive (runtime) security capabilities
- Specific focus on autonomous agent security, addressing an emerging threat vector
- Built-in red teaming capabilities for continuous security validation
While the enterprise AI security market is nascent, we expect rapid evolution as threats mature and regulatory requirements solidify. With Prisma AIRS, Palo Alto Networks establishes an early position. This continues the company’s impressive run of continuous innovation, adapting to emerging security challenges of enterprise AI. This is a strong offering and impressive leadership play by Palo Alto.
Competitive Outlook & Advice to IT Buyers
These sections are only available to NAND Research clients and it advisory members. Please reach out to info@nand-research.com to learn more.
