Agentic AI Is Redefining the Cyber Threat Landscape

Enterprises are entering a new era of cyber risk as agentic artificial intelligence begins to power attacks that unfold at machine speed. According to Carl Bryant, Managing Director for Cyber Risk Management at Unit 42, Palo Alto Networks, fully autonomous AI agents are fundamentally reshaping how threat actors operate—and how organizations must defend themselves.

Agentic AI refers to systems that can independently plan and execute tasks on behalf of a user. In the wrong hands, these agents can be deployed in parallel to probe networks, exploit vulnerabilities, move laterally, and exfiltrate data within minutes rather than hours. Bryant warns that this capability dramatically lowers the barrier to entry for less sophisticated attackers, enabling them to launch complex, multi-stage attacks with minimal technical expertise.

“Attackers can now deploy agents to do the work for them,” Bryant said, adding that the same tools are increasingly accessible to both cybercriminals and enterprises. However, many organizations still lack a clear understanding of how to securely deploy and govern AI systems.

To counter agentic threats, Bryant stresses that enterprises must be able to detect and respond at machine speed. This requires runtime security, continuous risk assessments, and automated defenses built specifically for AI-driven environments.