How Gemini finds and cites brands

Gemini can draw on what its models learned, and it can ground an answer in Google Search. If you want Gemini to find your pages at prompt time, your standing in Google Search matters.

How grounding works

In Google’s developer docs, grounding has a simple flow. The model reads the prompt and decides whether a search would improve the answer. If so, it writes and runs one or more search queries, works from the results, and returns an answer with citations.

Two things follow from that. Not every prompt triggers a search, and the model writes its own queries, so they may not match the words a person typed.

What Google-Extended controls

Google-Extended is a robots.txt token, not a separate crawler you can see in your logs. Google says publishers can use it to manage whether content Google crawls may be used to train future Gemini models, and for grounding in Gemini Apps and in Grounding with Google Search on Vertex AI.

Google also says it doesn’t affect a site’s inclusion in Google Search, and isn’t a ranking signal there. So blocking Google-Extended won’t remove you from Search, but it can limit how Gemini uses your pages. Weigh that before you set it.

What to check first

Confirm the pages you care about are indexed in Google Search.
Read your robots.txt for a Google-Extended rule, and make sure it says what you intend. The free AI crawler checker shows it.
Check how your brand reads on the pages Gemini cites in your category, including sites you don’t own.

Gemini isn’t AI Overviews

Gemini and the AI Overviews in Google Search are different products. A result in one tells you little about the other. See the Google AI Overviews guide.

How Ocita measures it

Ocita sends your prompts through Google’s API with Google Search grounding on, up to 5 times a week, then records mentions, position and citations. Grounding has no location setting, so Ocita passes your location in the instructions sent with each prompt. These API answers can differ from the Gemini app.