Find deviations and reopen milestones through MCP

You can use an AI assistant connected to AmpliFlow through MCP to find deviations involving you, check impact grading, read checklist templates and reopen project milestones. Your usual permissions and visibility rules still apply.

Improvements

Find deviations linked to your account

The assistant can filter deviations by responsible user, creator or involved user. It first checks which account the connection uses. A service account does not automatically identify the person behind it.

When you combine filters, a deviation must match all of them. To find deviations where you are either responsible or the creator, the assistant runs separate queries and combines the results without duplicates.

Propose and check impact grading

The assistant can retrieve your organization’s current categories and score definitions to propose impact grading based on the reported facts. Missing facts need clarification, and you confirm the complete proposal before it is saved.

After a confirmed change, the assistant can read back saved categories and scores. A selected category with a zero value is not yet graded. If a save result is uncertain, it needs inspection before another attempt; the change is not repeated automatically.

Read checklist templates without changing them

The assistant can find and read published, draft-only and archived checklist templates. The template’s state is shown, and reading does not publish, restore or change it.

Reopen a project milestone

The assistant can reopen a completed milestone in an active project without recreating it. The milestone’s name, due date, order and tasks are preserved, including which tasks are complete. Archived milestones are not restored automatically.

Older MCP connections may need to reconnect. Questions about finding deviations or managing templates and milestones through MCP? Contact support.

This page is also available in Swedish.