Cato Networks launches Agentic Threat Prevention to counter AI-assisted attacks

Cato Networks has unveiled Cato Agentic Threat Prevention, a new security feature that utilizes autonomous agents to predict the trajectory of an attacker within a customer's network. These agents generate and implement tailored defenses before the attack can gain traction. The system operates on Cato’s cloud-native platform, which already gathers network and security data from each customer.

The agents leverage this shared information. Enforcement is carried out at all Cato points of presence without the need for service chaining. The company positions this launch against the evolving nature of cyber threats. Recent advancements in artificial intelligence enable attackers to exploit vulnerabilities, weaponize them, and scan multiple environments in rapid succession, often completing an entire attack faster than it takes to investigate a single alert.

Shlomo Kramer, Cato’s CEO, emphasized that traditional security measures are insufficient against today’s AI-powered threats, stating that "prediction is the new prevention." The agents operate using the same telemetry Cato already collects, as well as the customer's own behavior and external threat intelligence. By mapping risk across users, applications, traffic patterns, and exposures, the software identifies potential attack paths.

Once the prediction phase is complete, the system crafts specific controls for that environment, rather than one-size-fits-all detection measures. This new capability builds upon Cato’s earlier Agentic CVE Mitigation service, which applies protections to newly disclosed vulnerabilities within 45 minutes. Combined with Cato’s Dynamic Prevention feature, which detects attacks that progress slowly over time, the two solutions together reduce the window between exposure and protection, addressing the full attack lifecycle from the moment a vulnerability is identified to the moment it is thwarted.