Identity security has entered its chaotic era. Human, non-human, and agentic AI identities no longer just coexist. They form an uncontrolled inheritance chain in which a human creates an agent, the agent spins up service principals, OAuth grants, and role assignments, and that whole chain keeps running long after the human changes roles or leaves. Most of these chains are being spawned by business users on low-code and enterprise AI platforms, outside traditional identity controls and largely invisible to security.
In this segment, Radiant Logic CEO Dr. John Pritchard joins us to unpack why this is no longer a visibility problem. It is an observability problem. And it is shifting the center of gravity in identity security from authentication to authorization. Listeners will leave with a clearer view of where their current IAM, IGA, and NHI programs fall short, and a practical lens for governing the rapidly expanding population of AI agents already inside their environments.
Segment Resources:
https://www.radiantlogic.com https://www.radiantlogic.com/resource/through-the-eyes-of-the-adversary/ https://www.radiantlogic.com/resource/identities-under-attack-how-adversaries-exploit/
To go deeper on what John discussed today, watch Radiant Logic's on-demand webinar Identities Under Attack: How Adversaries Exploit the Human-Machine-Agent Divide at https://securityweekly.com/radiantlogicidv.
- 0:00 - Introduction and “three identity problem” concept
- 0:36 - Human, non-human, and agentic identity distinction
- 01:30 - Why agents differ from traditional non-human identities
- 02:20 - Goal-driven behavior and unintended outcomes
- 02:51 - Identity shift from authentication to authorization
- 03:11 - Authorization and intent-based policy challenges
- 04:13 - Shadow AI and uncontrolled agent creation
- 05:07 - Orphaned and ownerless agents problem
- 05:46 - Context as the core security challenge
- 07:00 - Breaking down organizational data silos
- 08:03 - Identity as a “context system” problem
- 09:00 - Determinism vs non-determinism in agents
- 09:40 - Human-in-the-loop and AI-assisted policy decisions
- 10:47 - Defense-in-depth policy layers across systems
- 11:46 - Convergence of security functions (IAM, SOC, governance)
- 12:24 - Misunderstanding agents as “human equivalents”
- 13:15 - Need for agent specialization and policy differences
- 13:48 - CISO role, board-level alignment, and governance strategy
Dr. John Pritchard is Chief Executive Officer of Radiant Logic, the identity security platform trusted by 30% of the Fortune 500 and 60% of U.S. Cabinet-level agencies. He stepped into the CEO role in January 2025 after joining Radiant Logic in 2022 as Chief Product Officer, where he led the company’s evolution from a virtual directory provider into an AI-driven Identity Security Posture Management platform that unifies, observes, and acts on human, non-human, and agentic AI identities.
John is the architect of the Three Identity Problem framework, the uncontrolled inheritance chain that forms across human, non-human, and agentic AI identities in the modern enterprise. His work reframes identity security as an observability problem and is shifting the center of gravity from authentication to authorization, governance, and lifecycle across the full identity chain.
Prior to Radiant Logic, John held senior product and engineering roles at Okta, Adobe, and IBM. Under his leadership, Radiant Logic was recognized in three Gartner categories in 2025, including Identity Visibility and Intelligence Platforms and AI for Access Administration, and earned Gold in the 2026 Globee Cybersecurity Awards (IAM) and Silver in the 2026 Cybersecurity Excellence Awards (ISPM). He holds a Ph.D. in Innovation Management from the University of Denver and is a frequent Forbes contributor on identity security and zero trust architecture.
