AI Agents Drive Faster, Smarter Vulnerability Exploits

Zafran Security's Havdala on Exposure Management, Adaptive AI Workflows Rahul Neel Mani (@rneelmani) • December 9, 2025

Artificial intelligence-generated code expanded the volume of software flaws and intensified exposure across modern environments. Threat actors now use autonomous agents to scan the internet, identify weaknesses and deploy exploits within hours. This shift reshaped the speed and scale of risk for organizations that rely on traditional security tools.

See Also: A CISO’s Perspective on Scaling GenAI Securely

Snir Havdala, co-founder and chief product officer at Zafran Security, said attackers build specialized agents that handle reconnaissance, exploitation and rapid adaptation. These agents behave like distributed experts that refine techniques as conditions change. Legacy automation can't match this pace because static playbooks break when data, APIs or environments shift.

"You have to find the right expertise. It's really hard to hire, to find the experts. Experts who are using AI, they ...