The numbers that actually tell you whether ChatGPT, Gemini, and Perplexity are naming your brand, and which ones are vanity metrics dressed up as strategy.
Start with CITED's free AI visibility check to see your own mention rate, or read what AI visibility is first.
AI visibility metrics measure how often, and how favourably, AI assistants like ChatGPT, Gemini, and Perplexity mention your brand when people ask for recommendations in your category. The four core numbers are mention rate, share of voice, sentiment, and citation rate. Together they replace keyword rank as the scoreboard for AI search.
Rank position was a good proxy for visibility when search returned ten blue links and you could count which one was yours. AI assistants do not return a ranked list: they synthesise one answer from dozens of sources and name a short list of brands inside it, often three or four with no runner-up.
That means a page one Google ranking and a strong AI mention rate are two different scoreboards, scored by different mechanics. You can hold the top spot on Google for a term and still be invisible every time someone asks ChatGPT the equivalent question.
Most of what gets marketed as an "AI visibility metric" collapses into four measurements. Track these across a fixed set of buyer prompts and you have a real baseline instead of a one-off screenshot.
| Metric | What it measures | How to read it |
|---|---|---|
| Mention rate | The share of relevant buyer prompts where an AI assistant names your brand at all | Your baseline number. Track it monthly per engine, not as a single snapshot |
| Share of voice | Your mentions as a percentage of all brand mentions across the same prompt set | Shows whether you are gaining or losing ground to named competitors, not just whether you appear |
| Sentiment and framing | How you are described when you are named: as a leader, an alternative, a caution | A high mention rate with weak framing still costs you the buyer's shortlist |
| Citation rate | How often your own pages are used as a grounding source in answers that mention you | The strongest lead indicator that your content itself, not just your reputation, is doing the work |
There is no universal benchmark, because the number of competing brands and the specificity of your category both move the ceiling. A niche B2B tool competing against three named alternatives has a structurally higher achievable mention rate than a brand in a crowded consumer category with dozens of contenders.
What is consistent: most brands start near zero. On average, 92% of brands do not show up when their category is asked about in an AI assistant, which makes a modest, improving mention rate a real result, not a rounding error.
Track the trend, not the absolute number. A move from 10% to 25% mention rate over a quarter tells you your fixes are working. A single 40% reading with no history tells you almost nothing.
Manually tracking AI visibility metrics means writing down the real questions your buyers ask, running each one in a fresh chat (no memory, no history) across ChatGPT, Gemini, and Perplexity, and logging whether you are named, who is named instead, and how you are described. Repeat monthly on the same prompt set so the trend is comparable.
That works, but it does not scale much past a handful of prompts before it becomes a spreadsheet nobody keeps updating. CITED's free AI visibility checker automates the same process: enter your site, and it runs the buyer questions for you, scores your mention rate, and shows exactly which competitors are named in your place.
Run a free check and get your mention rate, share of voice, and sentiment in about 15 seconds.
Check my AI visibility →