AI Search Watcher measures your brand with three metrics: Position (where the AI answers place it), Visibility (how often they mention it) and the AI Search Score (both combined, from 0 to 100). Two more metrics, Citations and C. Rank, measure which websites the answers link to. Higher Visibility and Score are better. A lower Position is better.
Every metric covers the prompts, AI models and dates selected in the filters at the top of the page. Point at the info icon next to a metric to see its definition in the app.
Where to find the metrics
On the Brand dashboard, the AI Search Score card sits next to the Top 10 brands table. Below them are the Average brand visibility and Average brand position charts, then Top performing prompts and Top 10 cited domains in responses.
The Prompts page shows the same metrics for each prompt. Select a prompt to open its own page, with its top brands and top citations.
Position
Position is your brand's average place in the AI answers that mention it. 1 means the answers named your brand first. Lower is better.
Position counts only the prompts and runs where your brand appears. A brand mentioned in one prompt only, in second place, has a Position of 2. So always read Position together with Visibility. In the prompt lists, a dash means the AI did not mention your brand in that prompt.
On the Average brand position chart, a better (lower) position is drawn higher. An improvement shows as a positive change in green.
Visibility
Visibility is how often your brand appears in the AI answers, as a percentage. For one prompt, it is the share of that prompt's runs that mention your brand. On the dashboard, it is your brand's average across all your prompts. Higher is better.
Top performing prompts lists your prompts by your brand's results. Its Position, Visibility and Score are your brand's numbers in each prompt, the same as on the Prompts page.
AI Search Score
The AI Search Score is a number from 0 to 100. It measures your brand's visibility and position in the AI models you selected. The app labels the score by range:
Score | Label |
90 to 100 | Outstanding |
80 to 89 | Excellent |
70 to 79 | Good |
60 to 69 | Fair |
0 to 59 | Needs Improvement |
To hide the label, turn off Display AI Search Score label in Settings, under Display options.
The free AI Search Grader uses its own ranges and labels, so its score is not on this scale. See AI Search Grader: how it works and how to read your score.
What does a good result look like?
The app's own labels are the simplest guide: an AI Search Score of 70 or more is labeled Good or better. Compare your brand with the others in Top 10 brands, because they are measured on the same prompts in the same AI models.
Watch the trend over several checks rather than one week. AI answers change from one run to the next, so one check on its own can move up or down.
Citations and C. Rank
Citations is the number of times the AI answers cited a website as a source, across all runs. C. Rank is the website's average position among the citations. When you view several AI models, a model that did not cite the website counts one place below its last citation. On the Prompts page, # of Citations and Citation rank measure your own website's citations in each prompt.
Only ChatGPT Search, Google AI Overviews and Google AI Mode give citations. Your website always appears at the bottom of Top 10 cited domains in responses with its own rank, even if never cited. Which AI models are monitored, and where citations come from explains the rest.
Performance cards
Three cards name the AI model where your brand does best. Best performing in is the model with the highest AI Search Score. Best visibility in has the highest Visibility. Most cited in has the highest share of answers that cite your website.
The performance cards compare the AI models you track. With one model selected in the filter, all three cards name that model. To hide them on a monitor, point at the cards and select the X. You can turn them back on in Settings, under Display options.
Is brand perception a metric?
No. The brand perception card shows what one AI model says about your brand when asked directly. It does not change any metric. See Change a monitor's settings to rate it or turn it off.
How are the metrics calculated?
Each AI model you selected answers every prompt in the monitor. AI Search Watcher checks every prompt twice on every weekly update, and the metrics combine those runs. The definitions above are the app's own. The exact formulas are not published.
If a number looks wrong for a specific prompt, contact us through the live chat or at [email protected]. Send the monitor's brand, the prompt and the date of the check.
Next step
To see which prompts drive your numbers, open the Prompts page. Add, organize and export prompts shows how to group them with tags.