Skip to main content

AI Agent

Updated over a week ago

How does the AI agent work?

When a ticket arrives, Junto's AI agent analyzes it, gathers relevant context from your integrations, and either suggests actions or takes them automatically—depending on your settings. You stay in control of what the agent can do.

What's the difference between Semi-Auto and Full-Auto mode?

  • Semi-Auto: The agent suggests actions and waits for your approval before executing

  • Full-Auto: The agent works autonomously without requiring approval for each action

Can I control what the AI agent does?

Absolutely. In Settings → AI → Toolsets, you can:

  • Enable or disable specific tools

  • Set approval requirements (Always, Once Per Session, Never)

  • Assign risk levels to different actions

What are risk levels?

  • Low Risk – Read-only actions like viewing ticket details

  • Medium Risk – Actions that modify data like updating ticket status

  • High Risk – Actions with significant impact like running scripts

Can I customize how the AI agent behaves?

Yes. Add custom instructions in Settings → AI → Configuration to guide how the agent responds to tickets and handles specific situations.

Did this answer your question?