At the RSA Conference in San Francisco, Pentera CEO Amitai Ratzon outlined a cybersecurity landscape being rapidly reshaped by AI, where both attackers and defenders increasingly rely on machine-speed operations. He emphasized that while AI does not inherently advantage attackers over defenders, it amplifies the long-standing asymmetry in cybersecurity—attackers only need to succeed once, while defenders must succeed every time.
Against this backdrop, Pentera is evolving its platform with an AI-enhanced product architecture that blends deterministic and probabilistic approaches, enabling enterprises to safely simulate real-world attacks, generate tailored insights instantly, and deploy continuous validation and remediation at scale.
Core Takeaways
AI-Driven Cyber Warfare: Cybersecurity is entering an era of machines fighting machines, where both attackers and defenders operate at AI speed, increasing the intensity and scale of threats.
Deterministic + Probabilistic Approach: Pentera combines controlled, deterministic attack methods with AI-driven probabilistic scenarios to deliver advanced testing while maintaining safety and accountability.
Safe, Automated Validation: Automated penetration testing platforms can safely run continuously without human error, providing always-on validation across hybrid environments without disrupting operations.
Unified Remediation Lifecycle: With the addition of Pentera Resolve, organizations can consolidate findings, prioritize fixes, and automate remediation workflows across multiple security tools.
Key Quotes
Closing the Remediation Gap
“Customers have been telling us… ‘You’re tackling everything we need… but then the findings… just pile up and we have more and more things to remediate. How are you helping us?’”
“So whenever a new thing that needs to be patched comes up… everything goes on a remediation life cycle—we do the de-duplication, triaging, assigning ownership… so remediation takes care of itself.”
Automation as the New Security Standard
“We’ve developed an automated platform that’s safe by design, we’ve actually engineered the platform to run safely in live production environments without disrupting operations… the system delivers consistent, repeatable results without the variability of manual testing.”
“Pentera runs twenty four seven… when you’re with your kids or grandkids, Pentera runs automatically. And when you wake up in the morning, it shows you what happened when you were sleeping.”
Balancing AI Power with Safety
“What we’re doing… is the combination of LLM-based probabilistic attack scenarios with the deterministic approach, meaning we’re going to show our customers which parts of the attack… is LLM and AI-based, and which are deterministic.”
“Eventually, customers are afraid of LLMs going crazy, unguarded… so with the combination, we’re basically mitigating the risk and continuing to take the responsibility.”
