Managed AI Runtime security resource is an instance of the Cloud NGFW service platform, you can author and enforce network security policies similar to other Cloud NGFW resources in Strata Cloud Manager.

You can configure, author, and enforce Managed AIRS (AI Access Security) policies for your Managed AIRS using the Strata Cloud Manager (SCM) console. The workflow begins by organizing your resources into designated management folders in Strata Cloud Manager. You can then target a specific folder scope in SCM to set up AI Security Policies and optionally apply the predefined AIRS-Best-Practice configuration snippet to quickly baseline security settings.

To enforce targeted protection, administrators must create custom AI security profiles and set up model groups mapped to their specific AI environments, such as AWS Bedrock. These profiles define action policies to alert on or block threats like prompt injections, data leaks, and toxic content, with an option to enable custom model support for cloud-based inspection. Finally, an outbound decryption rule with a trusted certificate must be configured and attached to the policy so the firewall can inspect encrypted payloads and prompt strings.

For more information, see Author and Enforce Managed AIRS for AWS Policies in Strata Cloud Manager.