AegisAI Debuts Proteus to Automate Deep-Dive Email Security

Modern phishing attacks now evolve faster than behavioral systems can adapt. While legacy security stacks rely on historical signatures or anomaly detection, these methods fail against AI-generated spear phishing that mimics legitimate correspondence. AegisAI claims its new model, Proteus, solves this by acting as a virtual security analyst. Instead of matching text patterns, the system verifies the integrity of links, sender history, and routing infrastructure against signals beyond an attacker’s control.

Benchmark results underscore the shift in performance. In a test of 3,020 human-reviewed phishing attempts, Proteus identified over 99% of threats. By comparison, leading general-purpose models struggled, missing roughly 30 times more attacks. Beyond accuracy, the system operates at higher speeds, processing verdicts up to 7.2 times faster than standard closed-source models. This efficiency is critical for inline security, where messages are held until the model clears them for delivery.

The urgency for this technology stems from a sharp rise in automated threats. AegisAI reported that AI-generated spear phishing jumped from less than 1% of tracked traffic to over 20% during 2025. With attackers leveraging personalized, AI-written messages that achieve click-through rates as high as 54%, the company argues that general-purpose models are insufficient. According to CEO Cy Khormaee, the goal is to shift email defense from reactive filtering to evidence-based verification, ensuring that every message is judged on its own merits rather than its resemblance to past campaigns.