Cyware CPO Sachin Jade on Rethinking Detection Engineering for AI-Driven Attacks

AI agents are changing the economics and speed of cyberattacks, challenging many of the assumptions traditional detection engineering was built around. As attackers gain the ability to rapidly generate new payloads, vary techniques, and operate at machine speed, relying heavily on known indicators and repeatable patterns is becoming less effective.

In a recent InfoSec Relations article, Cyware Chief Product Officer Sachin Jade joins security leaders to explore how detection strategies need to evolve as agentic AI becomes a greater part of the threat landscape.

Sachin argues that asking how to detect an “AI attack” misses the bigger challenge. Rather than trying to identify whether a human or an AI agent is behind an attack, defenders should focus on the behavioral chain and the adversary’s underlying objective. Attackers and their tooling may change, but behaviors associated with credential access, persistence, lateral movement, and exfiltration provide more durable signals for detection.

This shift also changes the role of traditional indicators. IOCs remain valuable for enrichment, retrospective hunting, and known infrastructure, but increasingly need to be combined with behavioral detection and environmental context to improve fidelity and support faster decisions.

As AI compresses the time defenders have to detect and respond, the future of detection engineering will depend on combining behavioral signals with the right context and automation, enabling security teams to identify what matters and act before machine-speed attacks can progress.