AI Agents Reshape Security: Intent Is Everything

At Proofpoint Protect in San Diego, security leaders declared the enterprise AI debate has shifted from deployment to governance. Attackers now use AI to craft flawless phishing lures and chain exploits at machine speed, while internal AI agents act as insiders with broad system access. Intent-based detection is emerging as the critical defense layer. Proofpoint revealed two new agentic security systems built on knowledge graphs, tiered intent models for its Nexus suite, and a Community Hyperloop that shares detections across all customers instantly. Anthropic expanded its Mythos Preview vulnerability-hunting model to defenders first. Proofpoint also reported nearly $2.5 billion in annual recurring revenue, growing close to 20%.