What happens when Google shows no AI Overview, and how is that different from a failure?
It is recorded as not triggered: a distinct, valid observation status that must be backed by a screenshot showing the page without an AI Overview. A failure is a technical exception. Neither enters the mention-rate denominator, but they mean different things: not triggered says Google gives no AI answer for that prompt; failed says the sample was not captured. Both appear as such in reports and the workspace.
Version 1.1 · Last updatedIt is recorded as not triggered. Google AI Overviews do not appear for every query; when the page loads normally but no AI Overview is shown, the task is marked not triggered, a distinct and valid observation status kept strictly separate from failed. Not triggered must be evidenced: the page screenshot is stored to prove that no AI Overview appeared at that moment. Neither status enters the mention-rate denominator, but their meanings are entirely different.
Not triggered versus failed
| Aspect | Not triggered | Failed |
|---|---|---|
| What happened | The Google page loaded normally but showed no AI Overview | A technical exception such as a timeout, CAPTCHA, expired session or UI change; no result obtained |
| Valid observation? | Yes: the platform behaved definitively by not producing an AI answer | No: the sample was not captured and says nothing about the platform |
| Evidence stored | Page screenshot, timestamp, region, language, platform entry point, prompt version; status set to not triggered | Exception type and timestamp, plus any captured screenshot |
| In the mention-rate denominator? | No, reported separately | No, counted as a runtime exception |
| In the exception rate? | No | Yes: exception rate = failed plus blocked divided by planned tasks |
| Next cycle | The same frozen prompt is asked again | The same frozen prompt is asked again; within a cycle technical retries reuse the task ID |
Why not triggered must be evidenced
No AI Overview and no capture both look empty in the results. Without a screenshot they cannot be told apart afterwards, which would let a vendor fold technical failures into not triggered to flatter the exception rate, or count not triggered as failures to explain thin data. Miaowa GEO stores the page screenshot, timestamp, region, language, platform entry point and prompt version for every not-triggered task, so a buyer can pull any such record and see the page as it was, with no AI Overview on it.
Not triggered is itself information
If a buyer question stays not triggered on Google cycle after cycle, the finding is that Google currently gives no AI answer for that kind of query: buyers see conventional results there, and for that prompt your AI visibility hinges on ChatGPT, Perplexity and the other platforms. Conversely, a prompt moving from not triggered to triggered is a change worth noticing in its own right. Treating these as failures and discarding them throws that information away.
How it appears in reports and the workspace
- The monitoring log separates valid observations from runtime exceptions per cycle; not triggered counts as a valid observation, failed and blocked as runtime exceptions.
- Every not-triggered task drills down to its stable task ID, sampling time, status and screenshot.
- The public sample at /en/report uses the same status vocabulary: failed and not_triggered are shown as such, never merged or hidden.
- Nothing is re-asked outside the cycle, so a not-triggered prompt is never retried until an overview happens to appear.
- Does not triggered affect the mention rate?
- Not directly, because it is outside the denominator. It does shrink that cycle's denominator for Google, so read the Google mention rate together with the not-triggered count.
- Do other platforms have a not-triggered status?
- Not triggered refers specifically to Google showing no AI Overview. When another platform explicitly declines to answer, that is recorded as refused, likewise a valid observation and likewise outside the mention-rate denominator.
- Can you retry until the AI Overview shows up?
- No. Each prompt runs once per platform by default, and not triggered is the true state of the page at that moment. Refreshing until an overview appears would break the like-for-like basis between cycles.
Last updated . The capabilities page and the pricing page are authoritative where they differ.
Help centre
Back to the index, or email contact@miaowageo.com.