Jan 20
Loading video player...
In this video, we walk through a real AI-powered penetration test using Aikido’s AI Pentest. You’ll see how agentic AI goes beyond traditional automated scanners by applying logic and context to uncover vulnerabilities, including business logic flaws that other tools often miss. We cover the full workflow: setting up and scoping a pentest with natural language, adding authentication and code context, running the test safely, observing AI agents in real time, reviewing critical findings, and auto-fixing issues with generated pull requests. We also look at the final pentest report and explain why AI pentesting enables faster, more effective application security. Chapters 00:00:00 Introduction to AI Pentesting 00:01:50 Launching an AI Pentest in Aikido 00:02:45 Scoping, Authentication, and Context Setup 00:06:30 Running the Pentest and Safety Controls 00:08:52 Watching AI Agents and Reviewing Findings 00:11:43 Autofix, Reporting, and Final Takeaways