External Agents: Making Console Available Wherever You Work

Today, we are excited to announce that the full capabilities of Console are now available directly in your Claude, ChatGPT, Codex and Cursor sessions via Console MCP.

Your agents can now interact with Console the same way your people do, inheriting your identity and your policies, with every action logged in Console for transparency.

With this, Console becomes the governance layer not just for your employees, but for the agents they build and work alongside.

Console meets you where you are

If you live in Codex, you can now pull user, request, workspace, and Playbook data directly from your Console workspace via direct API.

You can also work conversationally with Console Assistant to:

• Provision access

• Build and troubleshoot Playbooks

• Investigate requests

• Escalate approvals

Console knows who you are, understands your company's policies, and lets you keep working without leaving your Codex session. 

Here's a sample of what that looks like in practice:

Pull from Console: Ask Console from inside Claude, Cursor or Codex

• "What were the top ticket categories last month?" Console surfaces volume, resolution time, and patterns, ready to drop into a doc or deck.

• "Pull all device models we're managing." Use the list directly in the script you're writing.

• “List the last 30 days of requests and group them by request type.”

• “Show me all of my Knowledge Base articles related to Okta.”

Work with Assistant: Have a conversation with Console

• Add a new team member to an existing Google Group.

• Investigate a specific request and escalate to a human if necessary.

• Hit a permission wall mid-debug in Claude Code. Ask Console to unblock you. It checks policy and either grants upgrade access or routes to your manager, you stay in your coding session.

• Flag an employee departure while finalizing offboarding paperwork. Console kicks off the Playbook: deprovisioning, license reclamation, device return, all in parallel.

Agent to Agent: Console governs the agents you build

• Your DevOps agent hits a permission wall. Instead of failing, it routes through Console, which checks policy, grants scoped access, and logs the action for audit.

• Your coding agent needs to know which auth systems your company uses. Rather than giving it direct access to Okta, it queries Console and gets a safe, scoped answer back.

Agents have arrived. Console is ready for them

We’re watching it happen in real-time: the engineers, ops leads, and IT teams we work with aren't just using AI to move faster. They're actively building with it. They're spending their time in Claude, Cursor and Codex, and using agents in all aspects of their work.

Those agents become stronger when they have access to the other tools in our user ecosystems. They need to provision resources, escalate permissions, and pull information from sensitive systems.

Console is where your IT and organizational context lives, and it should be available anywhere work happens.

If you’re interested in connecting Console to your AI agent harness, or you’re thinking about what agent governance looks like at your company, we want to hear from you.