> ## Documentation Index
> Fetch the complete documentation index at: https://docs.surnex.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Google AI Overviews

> Which of your tracked keywords trigger an AI Overview, and whether you're cited in it.

**AI → Google AI Overviews** shows AI Overview data for every keyword you track. It's the only page in the AI section that needs no input — the data is captured during each [rank check](/tracking/overview).

## Stat cards

| Card               | What it counts                                |
| ------------------ | --------------------------------------------- |
| **Total Keywords** | Every tracked keyword                         |
| **AIO Triggered**  | Keywords where an AI Overview appeared        |
| **Domain Cited**   | Of those, how many cited your domain          |
| **Citation Rate**  | Domain Cited as a percentage of AIO Triggered |

Read them left to right. **AIO Triggered** against **Total Keywords** is your exposure — how much of your keyword set now returns an AI answer. **Citation Rate** is your performance within that exposure.

Both matter, and they move independently. Exposure rising while citation rate holds means AI answers are spreading through your market and you're keeping pace. Exposure rising while citation rate falls means you're losing ground.

## The table

| Column         | What it shows                                  |
| -------------- | ---------------------------------------------- |
| **Keyword**    | The tracked keyword                            |
| **Position**   | Your organic position for it                   |
| **Cited**      | Whether your domain appears in the AI Overview |
| **References** | How many sources the AI Overview cited         |

Expanding a row lists the **Referenced Sources** — the actual sources the AI Overview drew on.

## Reading it

**Rows where you rank well but aren't cited** are the priority. Ranking third while an AI Overview above you cites three other sites means the answer is being given without you, and your position is worth less traffic than it looks. These are also the most winnable: you already have the authority, so the problem is usually that the page doesn't answer the question directly enough to be extracted.

**Rows where you're cited but rank poorly** are the opposite and are good news. AI Overviews don't simply reproduce the top results — a clear answer can be cited from well down the page.

**The References count** tells you how contested the answer is. An overview citing two sources is nearly winner-take-all; one citing eight has room.

**Referenced Sources** is where the actionable detail is. Open a few rows for keywords you care about and look at what's being cited — often industry publications, documentation, or comparison sites rather than vendors. Getting mentioned on those sources can matter more than editing your own page.

## Improving citation

Common patterns among cited pages:

* The question is answered **directly and early**, not after several paragraphs of preamble
* The phrasing matches how people ask, not internal terminology
* Facts are stated plainly enough to be quoted in isolation
* Structure — headings, lists, tables — makes the relevant passage easy to locate

None of this conflicts with ranking well; it's largely the same discipline applied more strictly.

## Empty state

**No AI Overview data yet** — "Run a rank check to detect AI Overviews for your tracked keywords."

A new project shows this until its first rank check completes. If it persists after checks have run, no AI Overview appeared for any tracked keyword — which is itself a finding: AI answers aren't yet a factor for your keywords.

## Where else this appears

The same data drives the **AI Overview Impact** panel on [Tracking → Overview](/tracking/overview), the three summary cards on [Tracked keywords](/tracking/add-keywords), the **AI Cited** badge on individual keyword cards, and the **AI Overview Rate** card on the [dashboard](/dashboard/overview).