Cato Networks introduces agent-based threat prevention capabilities

Key Points

• Cato Networks introduced Agentic Threat Prevention, an AI-driven system that predicts attack paths and auto-deploys protections across its cloud-based SASE platform.
• The feature processes 4.5 million traffic signals per customer weekly, blocking attacks before they progress by applying 71,000+ targeted restrictions per account without manual intervention.
• It complements Cato’s Agentic CVE Mitigation, which patches vulnerabilities in as little as 45 minutes, reducing exposure windows for enterprise security teams.

What is changing

Cato’s new Agentic Threat Prevention uses AI agents to predict how attackers might move through enterprise systems. It analyzes network and security telemetry, like user roles, traffic patterns, and security events, to identify risky activity that looks harmless in isolation. For example, downloading a tool might be flagged as suspicious if combined with late-night access or an admin’s credentials.

Around 4.5 million traffic signals are processed per customer weekly, generating over 345,000 condition matches. The system auto-applies restrictions like blocking malicious tool downloads and tightening access controls based on each customer’s environment.

Why it matters

This matters most to enterprise IT and security teams defending against AI-assisted attacks like ransomware, lateral movement, and identity-based breaches. Traditional tools often react after an attack starts, but this system acts faster than humans can respond, automatically stopping threats before they escalate.

The practical takeaway is that organizations using Cato’s SASE platform can reduce their vulnerability exposure window and manual response workload. However, it complements, not replaces, existing protections like firewalls and IPS, as it is designed for multi-stage attacks rather than rapid “smash-and-grab” threats.

It’s worth noting that performance and effectiveness are still early-stage metrics, based on tests from mid-May, so enterprises should validate real-world impacts. Have you tested AI-driven prevention tools in your environment? Share your experiences below.

Read the original source.