Enterprise endpoints have transformed into dynamic execution environments, expanding the corporate attack surface to include user-deployed AI applications, open-source packages, browser extensions, local LLM models, and Model Context Protocols (MCPs). Traditional endpoint detection and response (EDR) tools lack visibility into these active autonomous agents, leaving organizations vulnerable to data exfiltration via prompt injection, unmanaged API permissions, and malicious cross-agent interactions.
To close this security gap, Palo Alto Networks introduces Koi Agentic Endpoint Security, an AI-native control plane engineered to govern autonomous AI agents and secure the modern software stack in real time. Powered by a risk engine, Koi shifts endpoint protection from passive inventory logging to continuous, runtime risk analysis across three pillars: See All AI Software, Assess Risks, and Control the AI Ecosystem. This architecture enables secure enterprise AI adoption at scale without degrading developer productivity. Koi is available as a standalone endpoint security solution and as an integrated core module within Cortex XDR® and Prisma® AIRS™.
