Enterprises are deploying copilots and autonomous agents faster than their security teams can see them. Gartner projects that 40% of enterprise applications will embed task-specific AI agents by 2026, up from less than 5% in 2025 — yet the questions most security tools ask haven’t changed. They catalog what agents exist and what they’re configured to do, while the real risk lives somewhere else entirely: in the data flowing through those agents at the moment they read, reason, retrieve, and send.
That blind spot is no longer theoretical. Nearly half of organizations are entirely blind to machine-to-machine traffic and cannot monitor their AI agents, and 68.8% of boards now worry about sensitive data leaking through AI prompts and models. Meanwhile, the architecture that protected us for decades — data at rest, in motion, in use — is collapsing, because a single agent workflow can pull from a file store, process inside a model, enrich via an external service, and distribute through a collaboration tool in seconds, with no human ever touching the data.
This session reframes AI agent security as a data problem, not a configuration problem. We’ll break down the “three-legged stool” of agent security — data, configuration, and identity — and show why the data leg is the one most teams are missing. You’ll leave understanding where exposures actually emerge across an agent’s lifecycle, why legacy DLP and posture tools can’t see them, and what contextual, entity-aware enforcement looks like when it follows your content across every channel and AI workflow. Viewers will:
* Discover where sensitive data exposure emerges across the full AI agent lifecycle.
* Understand why configuration tools, legacy DLP, and posture management miss critical agent-driven data risks.
* Learn how data, identity, and configuration work together through contextual, entity-aware enforcement across every AI workflow.
Speaker: Gidi Cohen, CEO and Co-Founder, Bonfy.AI
Speaker Bio:
Gidi Cohen is a serial entrepreneur and seasoned software executive with a deep passion for innovation and building world-class organizations. As CEO and Co-Founder of Bonfy.AI, he brings decades of experience in leveraging advanced software technologies—including AI/ML, analytics, modeling, and simulation—to solve complex enterprise challenges.
Widely recognized as a pioneer in cybersecurity risk management, Gidi has a unique track record of developing cutting-edge solutions that empower global enterprises to proactively manage and reduce their attack surface. He is a respected thought leader in the security analytics space and a sought-after speaker at industry conferences worldwide, where he shares insights on how sophisticated analytics and unprecedented network visibility can transform enterprise security.
Prior to Bonfy.AI, Gidi spent over a decade leading Skybox Security as CEO and Founder, where he was instrumental in helping security leaders prioritize and remediate vulnerabilities with precision and speed. His visionary leadership and technical expertise continue to shape the future of cybersecurity and enterprise resilience.
Gidi holds a BSc and MSc in Computer Science from Tel Aviv University.
Hosted by Gary Miliefsky, CISSP, Publisher of Cyber Defense Magazine
