Airlock Digital announces the availability of “Agentic Control” with the release of Airlock Digital v7.1, aiming to prevent unauthorized actions by AI agents at the endpoint. The product evaluates an agent’s requested behavior at runtime and blocks execution when it does not comply with configured policies.
Both outlets frame the update as an extension of Airlock Digital’s existing “Deny by Default” and Application Control approach, which has traditionally focused on what employees can run on devices. Help Net Security emphasizes that Agentic Control adds “Agentic Steering” capabilities for agent behavior and positions it as a way to extend Deny by Default Application Control to emerging AI-agent use cases. PR Newswire highlights policy-based control, including visibility into what supported agents attempt, applying policy before execution, and returning the decision to the agent.
Overall, the two reports describe the same core release: runtime policy enforcement for AI agents, with decisioning fed back to the agent to limit unauthorized behavior on endpoints.