04 use ai action buttons

Use AI Action Buttons

What this does

AI action buttons let users run contextual LoopIQ agents directly from the page they are working on. Instead of opening a separate AI console, users can run the right action from a release, work item, incident, test execution, certification, or other supported surface.

Why buttons may differ by page

LoopIQ decides which buttons to show using backend policy checks. A button may appear or disappear based on:

• the page you are on
• the selected organization and team
• your role and permissions
• your subscription
• the integrations connected to the organization
• whether the action requires approval
• whether the record has enough context for that action

This prevents the frontend from exposing actions that the user should not run.

Run an AI action

1. Open a supported LoopIQ record or workspace page.
2. Find the AI action button near the page header, action menu, or relevant section.
3. Select the action.
4. Review any prompt, confirmation, or required context.
5. Run the action.
6. Review the generated output before using it.
7. Save, attach, approve, or discard the output depending on the workflow.

Examples

On a release certification, an AI action may review readiness and missing evidence.

On a story, an AI action may draft acceptance criteria or identify implementation risks.

On an incident, an AI action may summarize impact, likely cause, and follow-up actions.

On a test execution, an AI action may summarize failures or suggest flaky-test triage.

Approval behavior

Read-only actions may return a recommendation immediately. Actions that would update records, create work, attach evidence, or affect release governance may require approval before the change is applied.

Troubleshooting

I do not see the button

Ask an administrator to check your role, subscription, organization settings, integration configuration, and page context.

The action returns no useful output

Confirm the record has enough data. For example, release governance actions work best when the release has linked work, tests, evidence, security findings, and certification checkpoints.

The action says access denied

Your role may not have permission to run that action, or the action may not be available in your subscription.

Related articles

• Use AI Agents and Agent Settings
• Configure and Run BYOA Agents
• Use AI Governance Actions on Release Certifications