Ron Gabrisko & Oege de Moor | Altimeter Capital

From AI Breakthrough to Enterprise-Ready Pentesting

AI penetration testing is moving from research to enterprise deployment, but success depends on more than powerful models. It requires validated results, production-ready systems, and a go-to-market strategy that earns the trust of both security practitioners and executives.

In this discussion, XBOW co-founder Oege de Moor joins Databricks CRO and XBOW board member Ron Gabrisko to explore how autonomous AI is reshaping offensive security. The conversation covers XBOW's journey from research project to the first AI system to reach the top of HackerOne's global leaderboard, demonstrating its ability to discover and validate real-world vulnerabilities across production applications rather than benchmark environments.

The speakers explain why enterprise AI agents succeed only when they are grounded in real-world workflows, supported by verification systems that eliminate false positives, and trained through continuous feedback. They also share examples of AI uncovering critical SQL injection flaws, business logic vulnerabilities, and novel exploitation techniques that would require significant manual effort to identify.

Beyond the technology, the discussion examines what it takes to build an enterprise AI company. Topics include scaling from technical adoption to executive buy-in, combining self-service products with enterprise sales, expanding from a single product into a security platform, and preparing organizations for a future where AI-powered attackers operate continuously. The central message is clear: as offensive AI accelerates, defenders will need autonomous, validated AI systems of their own to keep pace.