WORKING GROUP DISCUSSION · MODERATE & TEACH
As AI adoption moves from chatbots to copilots and autonomous agents, sensitive enterprise data now flows through prompts, responses, files, embeddings, AI memories, and downstream actions. This teaching session will help CISOs and cybersecurity leaders translate emerging AI security frameworks into a practical AI data security program. We will identify the common control themes across NIST, ISO/IEC 42001, OWASP, and related guidance, then provide a step-by-step approach for discovering sensitive AI exposure, governing access, securing files and prompts, monitoring AI usage, and preparing for agentic AI risk.
OUR GUEST TEACHER
Himanshu leads LightBeam.ai, a data security and privacy automation platform helping enterprises discover, classify, and protect sensitive data across cloud, SaaS, and AI surfaces. He brings two decades of experience building systems for security and risk teams operating at scale.
DATE
May 20, 2026
LOCATION
Atlanta, GA · Virtual
FORMAT
90 min · Moderate & Teach
HOSTED BY PHALYNX
SPONSORED BY LIGHTBEAM.AI
WHAT IS A WORKING GROUP DISCUSSION
A Working Group is a moderated, closed-door discussion held under Chatham House Rules, where security practitioners exchange real-world insights on a focused topic. Participants are prompted with questions and encouraged to share their own challenges and approaches. The sessions are designed purely for open knowledge-sharing, peer learning, and advancing the discipline together.
WHO ATTENDS
Each session brings together 15+ active security practitioners across virtual and on-site attendance, kept intentionally small to make real discussion possible. This room will be built for the people closest to the AI security problem: senior security executives from Fortune enterprises and mid-market organizations, lead architects and risk managers who own or influence tooling decisions, threat analysts shaping their teams' approach to emerging exposure, and peer influencers actively pressure-testing vendor claims and best practices.
*Note to attendees: we do not permit vendors and consultants. You must be an active cybersecurity practitioner with a company email.
WHY ATTEND
Translate NIST AI RMF, ISO 42001, and OWASP LLM guidance into a working AI data security program.
A step-by-step approach to discovering sensitive AI data, prompts, responses, and usage.
Prepare for agentic AI risk: copilots, agents, and downstream actions that touch sensitive data.
ABOUT SESSION TEACHER
Himanshu Shukla leads Lightbeam with a mission to simplify how organizations secure and govern sensitive data. A technologist and entrepreneur, he previously built AIOps and analytics products at Nutanix that scaled to $100 million in revenue. Earlier in his career, he held engineering and leadership roles at Cisco, Ericsson, and Novell, and founded Parallocity, where he developed advanced runtime detection systems. Himanshu holds multiple patents in distributed systems and resource modeling and is a graduate of the Indian Institute of Technology, Delhi.
LOCATION
Atlanta, GA
Virtual attendance available
FORMAT
Moderate & Teach
90 minutes
AUDIENCE
Security Leaders, Risk
Practitioners, Architects
