idGenius Introduces Matrix, a Relationship Based Security Platform for AI Agents

idGenius v2 Matrix is built upon the premise of a shared identity layer for AI Agents that enables business logic, governance and cost containment.

CHICAGO, IL, UNITED STATES, October 2, 2026 /EINPresswire.com/ -- idGenius today announced Matrix, version 2 of its AI agent identity security platform, and its establishment as an independent operating company. Anomalix will continue separately, providing advisory, implementation and managed services.

AI agents are already in production and writing code across enterprise applications and cloud platforms. They use service accounts, APIs, credentials, call tools, access data and delegate work. A flat access record may show what an agent is permitted to reach, but it cannot explain the full business context: who owns the agent, which project it serves, why it needs a tool, and what depends on it.

Matrix uses a relationship ontology to securely connect agents with their owners, organizations, business purposes, projects, credentials, tools, data, policies and other agents. idGenius then uses that context to govern actions as they happen, in real-time.

“Organizations are using outdated approaches and tools to secure AI Agents” said Mohammed Elkhatib, CEO of idGenius. “AI agents operate more like an entity than identity. Matrix connects authority to business purpose. Our harness gives agents the business, security and operational context it needs at runtime. Developers and security team no longer need to work in silos to build, secure and stop an agent if needed.”

The idGenius agent harness can be used by other agents to obtain policy decisions before taking actions. Matrix uses Open Policy Agent (OPA) and its Rego policy language to express and evaluate rules informed by the relationship ontology, then delivers those decisions through integrations at Model Context Protocol (MCP) tool boundaries. An agent’s proposed tool call can therefore be evaluated against its identity, owner, purpose, permitted resources and current risk before it proceeds.

Matrix also provides an agent kill switch. When an agent must be contained, idGenius can deny further actions at connected enforcement points while showing the owner, resources and dependent workflows affected by the decision. The goal is to give responders both a way to act and the context to understand the consequences.

“An agent can follow a technically valid path and still do something outside its business mandate,” said John Johnson, co-founder of idGenius. “The harness lets other agents ask for a decision at the point of action. The relationship model supplies the context behind that decision.”

Matrix brings together six capabilities: discover, learn, govern, enforce, observe and audit. It also connects agent activity and token usage to business purpose, helping enterprises identify unnecessary spend and measure changes in risk, capital expenditure and operating expense.

“AI agents create new chains of authority across tools, data and organizations,” said Ed Jackiowak, an early investor in Anthropic and a member of the idGenius board. “Enterprises need to see those chains and retain control when an agent’s behavior or purpose changes. That is the problem Matrix is built to address.”

As an independent operating company, idGenius will focus on developing and delivering Matrix. Anomalix will retain its separate identity for advisory consulting, implementation and managed services.

About idGenius
idGenius develops identity security for AI agents and other nonhuman identities. Matrix combines a relationship ontology, an agent harness and policy enforcement integrations to help enterprises discover agents, govern their actions, contain risk, observe activity and produce audit evidence.

Media contact
Michael Kerns
mkerns@idGenius.com
www.idGenius.com

Michael Kerns
idGenius Inc
mkerns@idGenius.com
+1 3128885544
Visit us on social media:
LinkedIn