Visit Lumos at Booth 115 on the main floor, or visit us at the Non-Human and AI Identity Pavilion at Booth NH04.
Stop by, catch a demo
Meet Lumos at Identiverse
Stop by our booth and see Lumos in action. Live demos throughout the day.
Book a Meeting
Skip the booth crowd and get a 30-minute working session with a Lumos identity expert, tailored to your stack and your goals for the year.
We're at Identiverse to compare notes with identity leaders on the four shifts redefining the category:
- Modernizing IGA for cloud and hybrid reality. Most IGA tools were built for static roles, predictable users, and on-prem directories. That world is gone and the migration shouldn't take years.
- Governing non-human identities at scale. In customer environments we've analyzed, more than 95% of NHIs have no assigned human owner. Without ownership, reviews can't run, lifecycle policies don't apply, and findings sit in backlogs.
- Preparing for AI agents as the next wave. The same governance backbone that holds for service accounts and machine identities today is the foundation for what comes next.
Two talks. Two theaters.
Defending AI Agents With AI Agents: A Practical Guide for a Post-Mythos World
More code shipped last year than in the previous decade combined. Every line is a new attack surface, and your IAM team didn't scale with it. Attackers are already shipping agents that exploit your identity sprawl, while your access reviews still run on quarterly cycles.
In this session, Andrej Safundzic maps exactly where human-speed IAM breaks under autonomous systems and shows what the shift to runtime identity control actually looks like in practice. You'll learn:
- Why identity is moving from provisioning to the moment of execution
- How to build identity agents with the right goal, context, tools, and guardrails
- Where to plot your program on the Cyber AI Transformation curve—and what to prioritize next
- How leading teams are using agents to govern agents, in production, today
Not All NHIs Are Created Equal: Why Service Accounts, API Keys and AI Agents Need Different Controls
Humans wait days for access and keep it forever. Machines scatter across systems with no clear owner. Agents execute, chain decisions, and generate access paths no one explicitly approved. Three identity types, three broken models, and most IAM programs were only built for one of them.
In this session, Andrej Safundzic breaks down exactly how each identity type exposes a different failure in today's IAM model and what it takes to govern all three at once. You'll learn:
- Why provisioning and governance are the wrong control points for the modern identity landscape
- Where risk hides across each identity type: dynamic access, missing ownership, and self-directed workflows
- Why visibility without enforcement is just lagging insight
- How to move to a runtime control model that acts in real time
AI isn’t just another identity problem. It requires a new control plane and a shift in how teams approach it.
Panel: The New NHI Risk Reality - When AI Agents Start Making Decisions
Panelists:
- Sriram Santhanam, Chief Cyber Officer, Human Managed
- Andrej Safundzic, CEO, Lumos
- Don D'Souza, Director of CyberSecurity, Fannie Mae
