Radiant Logic has added agentic AI features to its identity platform to track and govern AI agents across different software environments.
The update extends the company's Identity Visibility and Intelligence Platform into a fast-growing area of security as businesses deploy more autonomous software agents across cloud and internal systems.
The new functions create an inventory of AI agents used across an organisation, link those agents to human owners and associated workloads, and monitor changes in behaviour, access and configuration. The system also recalculates risk scores for individual agents in real time and can trigger remediation steps when conditions change.
The announcement reflects a broader shift in identity security, as companies try to manage not only employees and service accounts but also software agents that can act with increasing autonomy. Security teams have raised concerns that these agents may retain permissions and continue operating after the employee who created or managed them has moved roles or left the business.
Radiant Logic described this as a "Three-Identity Problem", covering employees, digital workloads and AI agents. It said the links between those identities can create gaps in accountability when agent ownership is unclear or outdated.
Another issue is fragmentation. Major cloud providers and software platforms are building their own agent registries, but these tools do not typically provide a unified view across rival systems. That can leave companies with multiple records of agents, owners and entitlements spread across separate environments.
Its latest release aims to provide a single record of AI agents across platforms without requiring customers to replace existing identity tools. The system can federate across several agentic AI frameworks and connect with current identity, governance and access management programmes through the Shared Signals Framework.
The product update centres on three areas: visibility, posture assessment and cross-platform coverage. In practice, that means cataloguing agents and their dependencies, checking for stale or orphaned agents, assessing whether privileges match an agent's purpose, and producing a risk score that changes as ownership, behaviour or access patterns shift.
Chief Executive Officer Dr. John Pritchard linked the launch to earlier efforts to bring together fragmented identity data from multiple user directories.
"In the traditional identity environment, enterprises had dozens of user directories and no unified view, and we built the platform that solved it," said Dr. John Pritchard, Chief Executive Officer, Radiant Logic.
"Today, these environments have a handful of fragmented agentic registries and the same problem. With 80% of Fortune 500 actively deploying AI agents, Radiant Logic will do it again - unify the registries and govern the identities, on the same vendor-agnostic foundation that solved it before," Pritchard said.
The company is positioning the release as an extension of its existing "Unify → Observe → Act" model, which it has previously applied to identity data more broadly. That approach now covers AI agents by pulling together records, observing behaviour and risk signals, and enabling enforcement or remediation through connected systems.
Chief Product Officer Sebastien Faivre said customers did not want a separate identity system for AI agents.
"Every CISO I talk to is being asked to govern AI agents now, and none of them can afford to deploy a second identity stack to do it. We built this as a native extension of the Identity Data Platform," said Sebastien Faivre, Chief Product Officer, Radiant Logic.
"The same federation architecture already trusted by a third of the Fortune 100 now extends to agents across every major framework. One authoritative record across humans, non-human identities, and agents. No rip-and-replace, no new silo," Faivre said.
Radiant Logic said its technology is used by one-third of the Fortune 100 and by 60% of US Federal Cabinet agencies. The company has built its business around consolidating fragmented identity data from legacy systems, cloud services and machine identities into a single operating view for security and access teams.
The latest release shows how identity vendors are adapting to the spread of generative AI and agent-based software inside large organisations. As businesses give those systems access to data, tools and workflows, questions about who owns an agent, what it can access and how that access changes over time are becoming part of mainstream identity governance.
Radiant Logic said the system's operating principle for agent security is: AI recommends, humans approve, systems enforce.