Q&A: Radiant Logic CEO John Pritchard on Identity, Context, and the Next Phase of Enterprise AI Governance

As organizations accelerate AI, context becomes just as important as innovation. The rapid rise of low-code AI development, autonomous agents, and non-technical users building AI-powered workflows is creating new operational challenges that traditional security and governance models were never designed to address.

Radiant Logic CEO John Pritchard talked with Tech Channels about how AI is reshaping enterprise identity management, why context is emerging as the next critical capability for cybersecurity, and why organizations must move beyond simple visibility toward understanding the relationships between people, agents, systems, and data. Pritchard argues that the future of enterprise security will depend on connecting these disparate pieces into a cohesive operating model.

Q. How has enterprise AI adoption evolved over the past year?

A. I think we’ve done the opposite thing from last year, where most companies said, 'Go experiment super fast, because we don't want to miss this window.’ And now they're living with that decision, where their people did all this experimentation, and it's a little bit out of control. The fact that there are these two conferences [Black Hat and Ai4 2026 in Las Vegas during the same week] happening at the same time, is a personification of what's going on. There's this massive movement to try to capitalize on this disruptive thing, and an existential freakout about how we got to govern this thing that's happening.

Q. What operational challenges are organizations experiencing as AI adoption accelerates?

A. I'm 100% seeing OpEx unexpected spend clawback because they didn't budget for this type of AI spend. Some of my big enterprises are now putting a per headcount spend by employee just to try to rein things in. This is an adoption they didn't expect.

Q. How is AI changing enterprise identity management?

A. We've been spending a lot of time on what I've been describing as a three-identity problem. We already had a lot of challenges getting human and non-human stuff under control. Now, this new third identity type, agents, has shown up. Actually, a bigger issue that I'm seeing, at least in the enterprise space, is that all an organization’s non-IT staff are making things—everybody's now a developer. And these actions are taken outside of classic governance tools.

Q. Why are AI agents creating new governance challenges?

A. Organizations want to make it super easy for a non-tech person to go in and create an agent that does something. That agent ends up creating a couple non-human identities behind the scenes. If that person leaves, chances are none of that gets deprovisioned. So, a sprawl inheritance issue is developing. And it's a bit uncontrolled.

Q. Why is context becoming more important than visibility alone?

A. Context is the word I'm talking about more often than not. We have the same problem in cyber. We don't see all the context because it doesn't sit in one place. The role of context and how to connect the dots between all the different sort of layers is the next area we're going to have to figure out as an industry.

Q. What does the industry still need to solve?

A. We have the pieces. The endpoint people have a piece. The network people have a piece. The governance people have a piece. What we're not really good at is connecting those dots.

Q. Why is connecting data across security tools still so difficult?

A. Right now, you have to say, ‘Hey[so-and-so], tell me what you've got.' Someone's got to try to assemble that. There are a lot of issues with data.

Q. Has the industry become too focused on visibility?

A. I feel like we're getting a little over-rotated on visibility. Everyone's saying we have a visibility problem. I think we made a lot of strides. I think we have a causality problem right now. We don't understand the connections between things deeply. That's why I think the context issue is going to reach a tipping point.

Q. Do traditional behavioral analytics still apply to AI agents?

A. I'd say a qualified yes. All our insider threat stuff we've done for years, I don't think exactly applies. We don't have the human intent side that we used to model. What's abnormal now? The behavior, yes. And the modeling is different. We now must look at what's abnormal behavior in the objective world.

Q. How are buying decisions changing as AI and identity converge?

A. Traditionally the identity organization is sort of my champion. Many of those teams have gotten moved into a security organization. The teams are having to cooperate a bit more and creating what I call buying councils. “OK, I need you, you, and you all to sit down so we can talk about how this works.”

Q. Who else is becoming involved in security decisions?

A. The CIO, CTO, CSO—it’s a fluid buying council. Some of the data owners are becoming an interesting player. Data security is getting really close to how we do cyber.

Q. Why is observability becoming increasingly important in identity security?

A. The last area that I'm personally spending some time on is observability within identity. We did this in infrastructure for years. With this three-identity problem: human, non-human, agentic, connecting the dots. I think that becomes super important about how we understand what's going on. It’s super powerful to know why something happened.

Pritchard paints a clear picture of the next phase of enterprise AI governance. While organizations have made significant progress in gaining visibility into AI adoption, visibility alone is no longer enough. The greater challenge lies in understanding the relationships between people, non-human identities, autonomous agents, data, and enterprise systems. As AI becomes easier for employees to build and deploy, organizations must shift from simply discovering assets to establishing context, connecting information across disparate security tools, and governing increasingly complex identity ecosystems.

Throughout the discussion, one message remains consistent: the future of cybersecurity depends less on collecting more information and more on understanding why systems behave the way they do. Identity, observability, context, and connected governance will become the foundational capabilities that allow organizations to securely scale AI while maintaining trust, accountability, and operational control.