Agentic AI in Cybersecurity: Autonomous Threat Hunting and Remediation

ABSTRACT

The rapid escalation of automated cyber attacks has outpaced human response times, necessitating a shift toward autonomous defense mechanisms. This paper introduces an Agentic AI framework for continuous, autonomous threat hunting and remediation within enterprise networks. By deploying a swarm of specialized 'Hunter' and 'Patcher' agents equipped with constrained execution environments, the system autonomously identifies zero-day vulnerabilities, synthesizes patches, and deploys them without human intervention. The study demonstrates a 94% reduction in mean-time-to-remediate (MTTR) against simulated ransomware campaigns, fundamentally altering the economics of network defense.