Skip to main content
Start with the prompt’s status and answer history. A saved prompt, a queued request, a completed answer, and completed analysis are different states. The prompt may be configured correctly without having collected an answer yet.

​
Check the common causes

  1. Confirm the workspace, prompt filters, platform, and date window.
  2. Check whether the prompt is active, paused, archived, or set to Manual.
  3. Review the schedule’s time zone and next run.
  4. Check platform access and the workspace’s available allowance.
  5. Open the answer history and read the latest status or error.
If collection is already queued or running, avoid repeatedly submitting it. If the answer exists but analysis is pending, its metrics and sentiment may not yet be available.

​
Retry a failed collection

Read the failure first. Resolve any access or allowance issue, then use Run now for a fresh answer. A current run does not recreate the historical AI response that would have appeared at an earlier date. Some search experiences may not return an AI answer for every question. Check the result’s status and evidence rather than assuming every search must contain an AI Overview.

​
Contact support with a useful reference

Include the workspace name, prompt question or identifier, platform, collection date, and visible error. Send the answer reference when one exists. Do not include passwords, API keys, or magic-link URLs. See scheduling and usage for configuration help.