Wiz brings AI agents to its US government cloud security product

Cloud security vendor Wiz announced on 14 October 2026 that its AI agents are now available in Wiz for Gov, the version of its platform aimed at United States public sector customers. In a post on its own blog, the company said the agents "elevate traditional CNAPP" and are intended to help public sector security teams "accelerate security and innovation".

CNAPP stands for cloud-native application protection platform, a category of tool that scans cloud accounts, workloads and configurations for risk and groups the findings for security teams to act on. Wiz's framing is that software agents, which can carry out multi-step tasks with limited human direction, sit on top of that existing machinery rather than replacing it. The company described the goal as securing "the mission at machine speed", a reference to the pace at which government security operations are expected to run.

The announcement is tied to Wiz for Gov, which the company associates with FedRAMP, the US federal programme that authorises cloud services for government use. Agencies and their suppliers generally cannot adopt a cloud product for federal workloads until it has been through that process, so bringing a new capability into a government-facing edition is a distinct step from shipping it to commercial customers.

What the announcement does not say is as notable as what it does. The published material seen by The Agentic Times does not describe which tasks the agents perform, whether they can take action in a customer's cloud environment or only recommend changes, what approval controls sit in front of them, or how their activity is logged and reviewed. Nor does it state the FedRAMP impact level at which the agents are authorised, or whether the agent features were assessed separately from the rest of the platform. Wiz did not publish customer names or a general availability date for individual agent capabilities in the post.

Those details matter for buyers in regulated environments, because an agent that can read cloud configuration data is a different risk proposition from one that can change it. Government security teams evaluating the feature would need that information from the vendor directly.

The announcement is a vendor statement and has not been independently verified. Wiz has been approached for further detail on the agents' permissions model and authorisation scope.