AI search visibility metrics and KPIs tell you how often AI engines name your brand when buyers ask the questions you care about. You can't see real conversations, so every serious tool works the same way: a fixed set of prompts, sent to each engine on a schedule, with every answer stored and read. To see where you stand today, run a free AI visibility check on your domain, then read the metrics below.
What is AI visibility?
AI visibility is the share of sampled AI answers that mention your brand. A mention is your brand named in the text of the answer. A citation is different: it is a link to a page, shown as a source. You can be mentioned without being cited, and your site can be cited in an answer that recommends a competitor.
AI brand mentions are the raw material for every metric below. Each one is read from a stored answer, so you can always go back and check the sentence behind the number.
The six AI visibility KPIs at a glance
| KPI | What it counts | Formula | Good for |
|---|---|---|---|
| Visibility rate | Answers that name your brand | Answers with a mention ÷ all answers | The headline number, per engine |
| Average position | Your place in the answer when named (1 = first) | Sum of positions ÷ answers with a mention | Lists and "best X" prompts |
| Share of voice | Your mentions among all brand mentions | Your mentions ÷ all brand mentions | Comparing with competitors |
| Citation rate | Answers that link to your site | Answers citing your domain ÷ all answers | Earning links from AI answers |
| AI Overview trigger rate | Searches where Google showed an AI Overview | Searches with an overview ÷ searches run | Knowing where Google answers with AI |
| Sentiment | Tone of the sentence that mentions you | Positive, neutral or negative per mention | Spotting negative framing early |
Visibility rate
Visibility rate is mentions divided by answers. If ChatGPT answered your 30 prompts every day for 30 days (900 answers) and named you in 470 of them, your ChatGPT visibility is 52%. Report it per engine: a blended number hides where you are weak.
Average position
When an answer lists options, the order matters. Average position only counts answers where you appear, so always read it next to visibility: moving from 2.4 to 2.1 is good news only if visibility held.
Share of voice
Share of voice compares you with everyone else named in the same answers. To measure share of voice in generative AI engines, count every brand mention across the answers, then divide yours by the total. Example: 900 answers name brands 2,000 times; you are named 400 times, so your share of voice is 20%. It is the metric that moves when a competitor gains ground, even if your own visibility is flat. Share of voice in ChatGPT and share of voice in Perplexity are different numbers; keep them apart.
Citation rate
Search-backed answers attach sources. Citation rate is the share of answers that link to your domain. Pair it with the list of domains the engines cite most often for your prompts: those are the pages that shape the answer.
AI Overview trigger rate
Google shows an AI Overview for some searches only. Google says AI Overviews and AI Mode "may use a 'query fan-out' technique", issuing related searches to build the answer (Google Search Central). Track the share of your searches that show an overview, and treat "no AI Overview" as a valid result, not an error.
Sentiment
A mention in a negative sentence is not a win. Record the tone of the sentence about you and watch for answers that repeat an outdated complaint.
Free · 2 checks a day · results saved for 7 days
AI visibility metrics by engine
The six KPIs apply everywhere, but each engine needs its own reading.
ChatGPT
ChatGPT decides by itself when to search the web; when it does, it cites sources. Track visibility, position and citations separately for ChatGPT, and keep the method fixed: the ChatGPT app and OpenAI's API can answer differently. See the ChatGPT rank tracker.
Gemini
Gemini can ground its answers in Google Search; Google says this lets Gemini "cite verifiable sources beyond its knowledge cutoff" (Google AI for Developers). The useful AI visibility metrics for Gemini are visibility rate, average position and the sources it cites when it searches. Keep Gemini apart from Google's AI Overviews and AI Mode: same company, different products. See Gemini rank tracking.
Perplexity
Perplexity searches the web for its answers and lists sources, so citation rate matters most here, next to visibility. See the Perplexity visibility tracker.
Google AI Overviews and AI Mode
The trigger rate only exists here: an AI Overview appears for some searches and not others. Measure visibility on the searches that show one, and the pages they link. See AI Overviews tracking and the AI Mode tracker.
Claude, Grok and DeepSeek
Claude and Grok can search the web, so all six KPIs apply. DeepSeek is measured without web search in Brisano, so its answers come from the model alone: track visibility, position and share of voice, but not citation rate. See Claude, Grok and DeepSeek.
How to measure GEO success over time
What counts as success in GEO (generative engine optimization) is a trend, not a single number. Four rules keep the AI visibility tracking success metrics honest:
- Read trends, not runs. The same prompt can get a different answer an hour later. Weekly or 30-day windows are the useful unit.
- Keep the prompt set fixed. Adding or removing prompts changes the denominator. Compare periods with the same prompts.
- Split by engine. An engine that searches the web reacts to new pages in days; an answer from the model alone moves mostly when the model changes.
- Separate "not mentioned" from "no answer". A failed run or a search with no AI Overview is not a miss and should not count against you.
Then define success up front: visibility rate up over 30 days on the engines your buyers use, share of voice up against two or three named competitors, and a list of prompts won and lost, with the answer text.
GEO KPIs vs SEO KPIs
| SEO KPI | GEO KPI | What changes |
|---|---|---|
| Ranking position | Visibility rate and average position | AI answers have no fixed position, so you count how often you appear |
| Click-through rate | Share of voice | The answer is read on the page; who else is named matters |
| Backlinks | Citation rate and cited domains | The cited pages shape the answer, yours or not |
| Organic sessions | (from analytics, not from GEO metrics) | Traffic is measured where visits arrive |
SEO and GEO KPIs complement each other: SEO tells you how people find your pages, GEO tells you whether AI answers recommend you before anyone clicks.
What these metrics do not measure
None of these numbers is traffic or revenue. For AI search ROI, use the tools that see visits: Google says sites that appear in AI features are reported "in the Performance report, within the 'Web' search type" of Search Console (Google Search Central), and visits from AI apps appear as referrals in your analytics when the app sends a referrer. AI visibility metrics tell you whether you are recommended; analytics tell you who arrived.
An AI visibility report template
- Visibility rate per engine, last 30 days, with the change against the previous 30.
- Share of voice against your two or three main competitors.
- Prompts you lost or won this period, with the answer text.
- Top cited domains that mention competitors but not you: your next outreach list.
Brisano's Overview shows visibility per engine, position and share of voice for the same window, and the weekly report emails it on every plan; cited sources and citation gaps are in Growth and Agency. It is an AI visibility tracker: it measures the answers, it doesn't change them. Get started when you want this report without building it by hand.
Sources
- AI features and your website · Google Search Central, checked 2026-11-18
- Grounding with Google Search · Google AI for Developers, checked 2026-11-18
- Overview of OpenAI crawlers · OpenAI, checked 2026-11-18