In brief
This glossary explains every AI visibility number in GEO Monitor in plain English, from Visibility Score and mention rate to share of voice, prominence and sentiment. Each entry says what the number counts and how to read it.
Key points
- An answer is one engine's reply to one prompt. Most figures count answers.
- Headline figures use the last 7 days, with each prompt counted once per engine.
- Visibility Score is 50% mention rate, 30% citation rate and 20% positive sentiment.
- Share of voice compares your mentions with the competitors tracked on your prompts.
- Position-weighted figures count being named first more than being named last.
- AI answers vary from run to run, so read every figure as a trend.
This page explains each AI visibility number you will see in GEO Monitor. For each one you get what it counts, how it is worked out and how to read it. If you want to see these numbers in place, start with the Dashboard.
A few words first
Prompt, run and answer
A prompt is a question you track, such as "best kitchen fitter in Leeds". A run is one prompt being asked on its engines at one time. An answer is one engine's reply in that run. Most figures below count answers.
Which answers count
The headline figures are a snapshot of now. They use the last 7 days, and each prompt counts once per engine, using its latest answer. If you run a prompt again on one engine, its new answer replaces the old one rather than being added on top.
If an engine's next run is a day or two late, its latest answer is kept in the snapshot for a few extra days, so the figures don't dip while you wait.
Answers from market runs count separately for each market.
Read figures as trends
AI engines don't give the same answer every time. One week's figures are a sample. Watch the direction over several weeks, and use the dashed lines on the trend charts to spot dates when an engine changed how it answers.
Headline figures
Visibility Score
A single score from 0 to 100 that blends three signals. Half of it is your mention rate, 30% is your citation rate and 20% is your positive sentiment share. So a 40% mention rate, a 10% citation rate and 50% positive sentiment gives 20 + 3 + 10 = 33. The "/ 100" is the top of the scale. It has nothing to do with how many keywords you track. The trend chart plots the same score day by day.
Keyword Visibility
How many of your active prompts at least one engine named you for, shown as a count over your active prompts, such as 18 / 40, and as a percentage. A prompt counts once however many engines name you. Think of it as the "we appear for X of our N questions" figure.
Mention Rate
The share of answers that name your business. With 40 prompts on 4 engines there are 160 answers, so 40 mentions is a 25% mention rate. It is a per-answer figure, so it is usually lower than Keyword Visibility.
Citation Rate
The share of answers that cite a page on your own website as a source. Being named and being cited are different. An engine can recommend you without linking to you.
Engine mention rate
Shown in Engine Breakdown on the Dashboard. It is your mention rate on one engine, counted only from that engine's successful answers. The figure in brackets is mentions out of successful answers. "0 of 5 succeeded" means the engine was tried but no answer came back.
Sentiment
The tone of your answers, as Positive, Neutral and Negative shares that add up to 100%.
- Positive means you are recommended, praised or listed favourably.
- Neutral means you are mentioned factually, or the answer isn't really about you.
- Negative means you are criticised, warned against or compared unfavourably.
Each answer gets a quick tone rating when it comes in. Answers that name you are then checked again each morning by an AI model that judges how your business is portrayed. Answers that don't name you count as neutral, which is why neutral is often the largest share.
Competitor figures
Share of voice
Your share of all brand mentions, compared with the competitors tracked on your prompts. The app adds up how many answers name each brand, then works out each brand's share of the total. If you are named in 30 answers and competitors in 70, your share of voice is 30%. The app shows the top six brands. On the Dashboard, the "m" figure is the number of mentions and the "c" figure is the number of sources cited in answers where that brand appeared.
Position-weighted share of voice
The same idea, but a brand named early counts for more. In each answer, brands are put in order of the sentence where they are first named. The first gets 1 point, the second half a point, the third a third, the fourth a quarter and so on. Each brand's share is its points divided by all brands' points. You find it on the Competitors tab, where it is the default view.
Average position in answers
Where you are usually named among the brands in an answer. 1 means named first. 2.5 means you are typically second or third. It only counts answers that name you.
Prominence
A score from 0 to 100 for how strongly you show up when you are named. It combines four parts, each shown in the Prominence breakdown.
- Named early is how near the top of the brand list you are, from the position weighting above. It carries the most weight.
- Recommended is how often the AI recommends you rather than just listing you, from the Intelligence analysis.
- Cited is your citation rate.
- Positive tone is your positive sentiment share.
Named early counts for 30 points, Recommended 20, Cited 15 and Positive tone 10, scaled to 100. A part with no data yet is left out rather than counted as zero. The score falls into a band. 80 and above is Featured, 60 and above Visible, 40 and above Present, and 1 to 39 is Buried. A score of 0 shows as Absent.
Share of recommendation
From the Intelligence view on the Competitors tab. Of all the times the AI recommended a brand, the share that went to each one. It differs from share of voice because a brand that is only mentioned in passing doesn't count. Recommendation share by engine shows the same split for each engine. "rec 12x" means the brand was recommended 12 times.
Share of voice by market
On the Markets tab. Share of voice worked out separately for each market you have added, from your market runs over the last 90 days. Each card also shows how many responses it is based on.
Prompt and run figures
Visible
On the Prompts tab, the Visible column shows how many of a prompt's completed runs had at least one engine naming you, out of all its completed runs, such as 3/8. Live checks and market runs are not included.
Queries where brand is visible and Visibility rate
At the top of the Prompts tab. Queries where brand is visible counts the active prompts where at least one engine has named you. Visibility rate is that number as a share of the active prompts that have been run.
Visible on a run
On the Runs tab, "Visible 2/4" means two of the four engines that answered on that run named you.
pos.
Shown next to Visible on the Runs tab. It is the sentence in the answer where you are first named. pos. 1 means the opening sentence.
AI demand
On the Prompts tab. The estimated number of monthly searches in AI tools for a prompt's main phrase. "Low" means too few to measure, and the wider topic's monthly figure is shown underneath. Demand figures use credits.
Source figures
Top citation sources
The websites AI engines cite most in your answers, with how many times each was cited, such as 14x. Links to your Google Maps listing are listed separately. See Sources and AI crawlers.
Your URL cited
The number of answers that cite a page on your website. Click it on the Dashboard to see which answers and pages. See Cited pages.
AI Overview Coverage
From SEMrush, when it is connected. The share of your tracked keywords that bring up a Google AI Overview, such as 35% for 14 of 40 keywords.
Figures on other tabs
Some tabs have their own scores, explained in their own articles. See Answer accuracy for how AI engines describe your business, Topic gaps for topic coverage, and Google AI Overviews for AI Overview citations.