The AI Visibility Index
How ChatGPT, Claude, Gemini, Grok, and Perplexity decide which businesses to name, and how to measure it.
When a buyer asks an AI assistant for the best provider in a category, the model returns a short, confident list. That list is not random. It is the output of a retrieval and generation process that weighs how often an entity appears in training and retrieval data, how consistently it is described, and how strongly other sources corroborate it.
The AI Visibility Index measures the result of that process directly: it asks the questions buyers actually ask, to the models buyers actually use, and scores what comes back.
Six intents, five systems
Every scan runs the six canonical buyer intents, purchase, comparison, research, local, review, and emergency, against five systems in parallel: ChatGPT, Claude, Gemini, Grok, and Perplexity. Running the same intents across models exposes a fact that single-model checks miss: visibility is uneven. A business can be named consistently by one model and ignored by another.
What the score is made of
The composite AI Visibility Score combines three measured signals. Mention rate captures whether you appear at all. Recommendation strength captures whether you are named first or buried in a list. Sentiment captures the language used about you when you are named.
Three derived dimensions add depth. Machine Trust isolates sentiment quality among the answers that mention you. Consensus Strength measures how much the models agree, because a business named by one model and missed by four has a fragile presence. Citation Density measures how grounded the answers are in citable sources you can influence.
Why a single number is not enough
A score is a starting point, not a verdict. The per-model breakdown is where strategy lives. Disagreement between models points to thin or inconsistent source coverage. A high mention rate with weak recommendation strength means you are known but not preferred. The Index is designed to make those distinctions visible, not to flatten them.
Honesty about estimates
Two outputs are models, not measurements, and we label them as such. Revenue opportunity is an industry-calibrated estimate of demand exposed to invisibility, not a figure pulled from your books. Ranking position is inferred from answer structure. Everything else is a direct reading of what the models said, captured as evidence you can inspect line by line.
See where you stand.
Run a free, live AI Visibility Scan across five systems in under a minute.
Run the scan →