ThreatLocker Launches Zero Trust Control for AI Agents
The new tool lets managed service providers restrict AI agent data access across enterprise applications.
Updated on Oct. 21, 2026 in Cybersecurity
Live Poll
Do you trust that your workplace can effectively manage the security risks of using AI tools?
ThreatLocker has launched Zero Trust AI Control, a new security product that allows managed service providers to restrict data access for AI agents on corporate computers. This security layer operates independently of the AI system to enforce strict permission boundaries.
Why it matters
The tool addresses security vulnerabilities arising when AI agents interact with business applications that possess broader data access than the agents themselves require. It enables companies to integrate agentic AI while maintaining granular control over corporate data.
The system enforces security policies that follow AI agents across applications such as Excel and PowerShell. It mandates device verification before granting access to company-sanctioned AI accounts.
The players
ThreatLocker
A cybersecurity firm that provides zero trust endpoint protection, application allowlisting, and ringfencing solutions for managed service providers.
The details
The software functions by treating AI agents as distinct identities, allowing administrators to assign permissions that remain active as the agent interacts with different applications. It integrates with existing application allowlisting and web filtering, providing suggested policy configurations for customization. By separating the permission layer from the AI application itself, the tool prevents agents from accessing sensitive files or data beyond their defined scope.
Timeline
October 20, 2026: ThreatLocker officially launched the Zero Trust AI Control product.
The Tech Race
This launch represents an effort to standardize security for agentic AI within existing enterprise software stacks. It aims to catch up with the rapid deployment of autonomous agents by retrofitting traditional endpoint security controls onto modern AI workflows.
Managed service providers can now deploy this control layer to manage access for AI agents currently in use on corporate workstations. Organizations will need to configure these new policies alongside existing allowlisting and web filtering rules to ensure consistent coverage.
The takeaway
This product shifts the responsibility of AI security from the model provider to the enterprise infrastructure layer. Security teams should monitor how these policies hold up against complex multi-step AI agent workflows in the coming months.
Further reading
Learn more about the latest developments in Cybersecurity.
Live Poll
Do you trust that your workplace can effectively manage the security risks of using AI tools?
