TrustLogix Unveils AI Agent Kill Switch to Combat Growing ShadowAI Risks

As enterprises rapidly deploy autonomous AI agents across internal systems, cybersecurity teams are struggling to control how those agents access sensitive corporate data.

TrustLogix announced a major expansion of its TrustAI platform aimed at solving that problem with real-time AI governance controls at the data layer.

The new release introduces several capabilities designed for enterprise AI security, including a runtime “kill switch” that allows organizations to instantly revoke an AI agent’s access to enterprise data if it behaves suspiciously or exceeds its intended scope.

TrustLogix also launched what it calls Intent-Based Access Control (IBAC), which evaluates not only who is requesting access, but why an AI agent is requesting it. The company says the approach helps enforce least-privilege access for autonomous agents operating across multiple systems and workflows.

The update also includes a new MCP Data Gateway for securing Model Context Protocol environments, along with a “Guardian Agent” that continuously monitors AI agent behavior for anomalies, policy drift, and unauthorized ShadowAI activity.

“Every enterprise we talk to is asking the same question: we have IAM, we have an AI platform, so why do we still have a data access problem?” said Ganesh Kirti, founder and CTO of TrustLogix. “The answer is that neither was built for the data layer.”

The release comes as enterprises across healthcare, financial services, and technology race to deploy AI agents faster than governance teams can secure them, creating growing concern around data exposure, compliance, and unsanctioned internal AI deployments.