The Cybersecurity Defenders Podcast · LimaCharlie

#255 - Defender Fridays: Identity Automation in the Age of Agentic AI with Matthew Chiodi from Cerby

·33 min·3 clips
Why do 78% of leaders distrust autonomous solutions, and how does a human-in-the-loop model address this?
This episode of Defender Fridays features host Maxime Lamothe-Bressard in conversation with an industry expert on identity automation in cybersecurity. The discussion centers on the practical implementation and current limitations of autonomous security solutions. A key statistic cited is that 78% of leaders reportedly lack full trust in autonomous agentic systems. The guest explains their company holds a patent specifically designed to incorporate a human-in-the-loop model. They state the industry is still years away from deploying any fully autonomous security operations. While technically possible to run systems fully autonomously, the guest asserts their firm would never do so at this stage. The conversation explores the expanding defensive surface in cybersecurity that necessitates these tools. The format is a live, 30-minute fireside chat intended for hands-on practitioners. A surprising claim is the explicit prioritization of human oversight over full automation despite available technology. The guest frames their patented model as a direct response to widespread leadership skepticism. An interesting insight is the admission that full autonomy is feasible but deliberately avoided due to maturity concerns. The discussion implies a significant gap between technical capability and organizational readiness for trust. The tone is informal and conversational, mirroring an unfiltered discussion between peers. The style is exploratory, focusing on real-world challenges rather than presenting polished slides. Cybersecurity professionals interested in identity management and automation strategy would find this episode valuable. Listeners seeking highly technical, step-by-step tutorials might prefer more instructional content.
Listen to the show on