Jamf launches AI Governance for Mac in ANZ

Jamf has launched AI Governance in Australia and New Zealand, a capability within Jamf for Mac that gives IT and security teams visibility into AI tool usage across their device fleet, policy controls enforced at the endpoint, and audit-ready reporting.

The company said the release makes it the first vendor to deliver native, OS-level AI governance controls for Mac - a claim based on the absence of any existing tool that unifies platform-native device management, deep AI tool configuration and a workflow that translates governance intent into vendor-correct configuration on macOS, Jamf stated.

The capability launches with support for Claude Code, Claude Desktop and OpenAI Codex, covering model access, tenancy, network permissions, file system controls, MCP server restrictions and other vendor-specific AI configurations.

A vendor control tracking engine continuously monitors supported platforms for new or updated controls, and all policies are enforced offline and before a user's first login to an AI agent.

Jamf said the rationale for a native approach is that AI tools running on Apple Silicon operate as processes that existing network proxies and cloud-based tooling cannot fully observe or govern. AI Governance uses Jamf's existing telemetry agent, built on native macOS frameworks, meaning no new agent is required on managed devices.

Organisations can use the capability to discover shadow AI tools, apply access policy at scale across different teams, and generate executive AI posture reports for CIOs and CISOs. The capability is designed for compatibility with SIEM tools and existing compliance frameworks, according to Jamf.

The release also includes an integration with Okta for AI Agents.

IT and security teams can register AI agents discovered on macOS devices directly with Okta, giving each a managed identity and scoped access to authorised cloud resources. Jamf controls which MCP servers can run on the device; Okta controls what cloud resources those servers can reach.

Agents use short-lived credentials rather than static keys, and the integration deploys from Jamf's console without manual API setup.

Organisations can also configure a preferred agent builder platform, such as Amazon Bedrock AgentCore, to route AI traffic through sanctioned cloud infrastructure.

"Organisations need governance that matches the way AI tools actually operate on Mac," said Beth Tschida, chief executive at Jamf.

"This means visibility into what's running, policy controls enforced directly on the endpoint, and reporting that helps security teams demonstrate compliance. Our AI Governance capability delivers that natively from the same platform customers already trust to manage and secure Apple devices."