A Conversation With Brian Vecci

https://ul.live/varonis_yt I had an awesome chat with Brian from Varonis about how AI is completely flipping the cybersecurity world upside down. We dive into why traditional security perimeters are failing and how to actually lock down your company's data when AI agents are running the show. What We Talk About: The AI Security Evolution: How enterprise conversations have quickly shifted from basic ChatGPT guardrails to managing complex, transformative AI agents. Offensive AI and New Attack Vectors: How new models are being used to automatically spot unpatched vulnerabilities, exploit them, and act as automated pen-testers. Why Perimeters Are Dead: Why the old "candy bar" defense of firewalls and endpoint protection fails when AI agents can bypass interfaces using API keys and inherited privileges. The "Trader" Copilot Story: A real-world example of a bank trader asking an AI copilot for investment trends and accidentally pulling up a massive, unsecured list of employee PII. A Holistic Defense Strategy: The importance of combining deep data observability, automated remediation, and behavioral monitoring to secure environments from the inside out. 00:00 - Introduction 00:35 - The shifting conversation around AI and security 02:20 - AI as an attack vector and automated pen testing 03:46 - The strategic approach to security and data observability 05:51 - Why perimeter defenses are failing against AI agents 11:37 - Real-world story: The bank trader and Microsoft Copilot 15:26 - Varonis's holistic solution and the launch of AI Atlas 20:38 - Deployment timelines and seeing immediate results 22:25 - Wrapping up and upcoming Blackhat announcements Subscribe to the newsletter at: https://danielmiessler.com/subscribe Join the UL community at: https://danielmiessler.com/upgrade Follow on X: https://x.com/danielmiessler Follow on LinkedIn: https://www.linkedin.com/in/danielmiessler/