Pentera Introduces Adversarial AI Agent To Guide Offensive Security Practitioners

Available in Pentera 8, Pentera’s AI turns complex offensive security testing into an interactive dialogue, dramatically increasing productivity

Boston, MA – March 19, 2026 –  Pentera, the Exposure Validation Company, unveiled Pentera 8, a major platform release introducing AI-powered attacks controlled in natural language. The release introduces Pentera Peer™, a natural language AI interface that guides users through their adversarial exposure testing, helping security teams understand exposures faster and take action with confidence.

Pentera Peer shortens the learning curve of adversarial security testing and transforms the testing process from a linear workflow into an interactive experience. With Pentera Peer, security teams can follow the attack testing progression, understand the implications of each step, and guide validation decisions on the fly.

Sample Pentera Peer prompts:

• Pentester: “Which test finding would validate privilege escalation paths toward domain admin, and which ones should I approve first?”
• CISO: “I’m preparing a briefing for the CEO and Board. Which validated attack paths currently expose our most critical systems and applications?”
• Remediation team leader: “Please create a remediation plan based on the exploitability and criticality of finding to reduce organizational risk based on a team of three.”
• Risk & Compliance Manager: “Based on our latest validation results and available threat intelligence, which exposures are most likely to be targeted and impact our PCI compliance posture?”

“Security teams need AI in their proactive testing programs, but they can’t rely solely on probabilistic methods that can, in some cases, introduce major risk,” said Amitai Ratzon, CEO of Pentera. “Pentera combines agentic AI that guides users through validation with AI-driven attack capabilities that adapt to complex environments, all grounded by a deterministic attack engine. The result is deeper exposure validation with the consistency, safety, and repeatability organizations need to test their security controls”.

Pentera Peer is an embedded agentic AI interface that allows users to extract insights from their testing results and guide adversarial testing using natural language. Built on Pentera’s attack engine and validated findings, Peer helps teams investigate exposures and interpret their impact based on organizational context such as role, industry, and geography. Teams can then confidently approve the next testing steps in real-time.

Pentera 8 will become generally available beginning Q2 2026, and will be demonstrated at RSA Conference 2026.