Live Webinar
We Set Claude Loose on Our Network: What Happened When an AI Attacker Met Microsegmentation
Amid a tidal wave of vulnerabilities and rogue AI agent anxiety, security teams face a practical defensive question: if an autonomous attacker compromises one machine, how far can it go?
To answer, we used the exclusive dual-use AI research capabilities granted as part of Zero Networks’ admission to the Anthropic Cyber Verification Program to run Claude Code as a live autonomous attacker against a Zero-segmented environment.
In this session, Rob Rachwald, VP of AI Security Strategy, and Nicholas DiCola, Chief AI Security & Customer Officer, will walk through the results of the experiment, outlining:
-
The techniques the agent used to escalate privileges, bypass controls, and move toward high-value systems
-
Why Claude only managed to compromise one intentionally exposed system, but could not move beyond it
-
Which defenses the agent disabled, and what the experiment’s results tell us about the role of microsegmentation in containing AI-speed attacks
Learn what happens when an autonomous attacker goes head-to-head with microsegmentation, and learn why proactive containment is the only reliable defense against AI threats.
"When I first saw Zero Networks, I walked away saying, this is too good to be true. When we put it in production, it was like a dream came true."
"My first impression was, it is too good to be true. The ease of deployment was shocking to me. It’s a simple and elegant solution to a very difficult problem."
"Zero Networks has become a core component of our security stack. It is something that we cannot live without any longer."
"Zero Networks is a game changer. The military-grade security it delivers across segmentation, access, and user rights, makes it a new cornerstone technology."
"I had to see it to believe it. The product was up and running in 15 minutes, it’s set it and forget it, and gives me peace of mind."
"When we ran pen tests before Zero Networks, attackers would get in within 5 minutes. Now, they cannot move laterally. It's a superb tool."
"We were having problems stopping some of the traffic in a pen test. Once we had Zero Networks in play, we were able to stop these activities."
"After [implementing] Zero Networks, we were able to lock down endpoints both on the server and on the laptops, our attack surface has decreased... incredibly."
"To do this by GPOs would have taken 300% more effort… we needed one FTE just to do this before Zero. Now, we can easily share the work."
"I would highlight the simplicity of [Zero Networks], both in the implementation and in the policy solutions that it generates for you. It's like they say in my department – it's black magic."
