Airlock Digital Introduces Agentic Control
In a groundbreaking announcement on November 17, 2026, Airlock Digital unveiled its latest security feature,
Agentic Control, designed to provide organizations with enhanced management capabilities over AI agents operating within their systems. As AI technology continues to evolve, the security landscape sees an increasing number of AI agents, which perform tasks autonomously and often carry the same access privileges as human employees. This rising complexity raises crucial questions regarding what actions these AI agents should be permitted to execute, and Agentic Control aims to provide clarity and security in this domain.
What is Agentic Control?
Agentic Control, a significant update included in the Airlock Digital v7.1 release, incorporates
Agentic Steering functionalities that empower users to apply specific policies governing the behavior of AI agents. The essence of this system lies in its ability to evaluate an AI agent's requested actions at runtime, blocking any unauthorized behavior before it can be executed. This proactive approach is particularly critical given that, according to a report from the Cloud Security Alliance in April 2026, 82% of organizations had unidentified AI agents operating within their infrastructure, with 65% experiencing incidents linked to these agents in the previous year.
How Does it Work?
Traditional application control systems typically focus on the software applications authorized for human users. However, with AI agents now dynamically engaging with systems, the need arises for new protocols that can both monitor and regulate these intelligent entities. Here’s how Agentic Control operates:
1.
Visibility Over AI Actions: Users can visualize all actions attempted by AI agents along with those that were blocked, providing a comprehensive overview of agent behavior.
2.
Policy Enforcement: Before any action is executed, the AI agent's request is assessed against pre-established customer policies, enabling IT teams to stop actions that deviate from approved uses while simultaneously allowing permitted activities to proceed.
3.
Steering Rather than Blocking: Rather than simply cutting off an agent's capabilities, which could lead to risky alternatives or escalation, the system communicates the denial of access, offers context, or suggests alternatives. This strategy allows the AI agent to learn and adapt its behavior within the approved operational boundaries.
4.
Continuous Improvement: Agentic Control maintains records of agent actions, decisions made, and their outcomes, enabling organizations to refine and adjust policies over time based on empirical data.
Bridging the AI Security Gap
David Cottingham, Co-founder and Chief Product Officer of Airlock Digital, emphasized the importance of managing AI agents effectively. He stated, “AI agents don’t simply halt when an action is blocked; they pursue alternatives, which can introduce risks. Agentic Control allows clear definitions of what AI agents can do, thereby maintaining safety while empowering them to act efficiently.”
Furthermore, Kevin Dunne, Chief Executive Officer of Airlock Digital, noted that the adoption of AI agents is advancing at a pace that often exceeds the capability of existing security measures. “Organizations require a clear strategy for monitoring AI agent activity, defining permissible actions, and ensuring compliance with organizational policies,” he remarked.
A Comprehensive Solution
Agentic Control isn't just an isolated solution but serves as a complementary tool to existing AI platform controls, identity management systems, and threat detection and response mechanisms. By instituting these preventive checks directly at the endpoint level, organizations can manage the actions that take place more effectively. The goal is to enable businesses to adopt AI technologies confidently without compromising security or operational integrity.
Available for Airlock Digital customers using v7.1 with Codex and Claude, Agentic Control was first showcased during Black Hat USA 2026, signaling Airlock Digital’s commitment to leading in the field of AI security. To explore more about this innovative solution, interested parties can visit
airlockdigital.com/agentic-ai-usage-control.
In summary, as AI agents become ubiquitous in modern business processes, having robust, policy-based control mechanisms at play like Agentic Control will be essential for organizations aiming to maintain security and operational efficiency.