05 use loopiq assistant with governed actions

Use LoopIQ Assistant with Governed Actions

What this does

The LoopIQ assistant can answer questions and, where enabled, help perform governed actions using LoopIQ context. The assistant backend decides whether a request is a normal question, an AI agent request, or an MCP-powered action.

The frontend does not directly expose raw MCP tools. The backend applies policy checks before any tool or action is used.

Before you begin

Make sure:

• your organization has assistant or AI features enabled
• your role can use the assistant
• you are in the correct organization and team context
• you understand whether the request is read-only or may create/update records

Use the assistant

1. Open the right-side LoopIQ assistant.
2. Ask a question or request help with the page you are viewing.
3. If the assistant suggests an action, review the proposed action carefully.
4. Confirm only if the action is appropriate.
5. Review the result and any linked records or generated output.

How LoopIQ governs assistant actions

For each request, LoopIQ can check:

• authenticated user identity
• selected organization and tenant
• user permissions
• subscription access
• page context
• allowed tool set for that page
• approval requirements

Write actions may require explicit confirmation or approval before they are applied.

Good use cases

Use the assistant to:

• summarize a release or work item
• find related evidence
• explain missing certification context
• locate security findings
• draft a follow-up comment
• prepare a remediation plan
• answer questions about visible LoopIQ records

Avoid these patterns

Do not ask the assistant to bypass permissions, mix data across organizations, use stale tokens, or make production-impacting changes without review.

Related articles

• Use LoopIQ MCP Tools
• Use AI Action Buttons
• Manage Roles, Permissions, and Access