What is stored for every answer?
- The full answer text, kept for good.
- Every brand it named, yours and your competitors', with the position each was named in.
- The pages it cited, with their rank where the engine gives one.
- The web searches the engine reported running to write it, and any follow-up questions it suggested to the reader.
- The engine, the model version, the country and the time it was asked.
Any number in the app opens to the answers it was counted from, in Response Details, so you can read what the engine actually said.
What does the first scan page show?
The First scan tab of AI Visibility reads the first scan in plain words, in this order:
- Does AI name your brand? How many answers to questions that do not name you mention you, by engine.
- Who AI recommends instead: the competitors named in those answers, and how often.
- The sources AI cites: the pages and domains the engines quoted.
- What the cited pages say: a summary and exact quotes from each source.
- Every question, every engine: one grid of named or not named.
- What the engines searched: the searches behind the answers.
- Is your site ready for AI? The site check's findings, such as whether AI crawlers can read your pages.
What metrics best measure brand visibility in AI answers?
The Visibility tab tracks four, over any date range, against the competitors you chose:
| Metric | What it counts |
|---|---|
| Mention rate | The share of answers to questions that do not name you in which you are named. The headline figure, with a 95% interval. |
| Share of voice | Mentions of your brand as a share of mentions of your brand and its tracked competitors. |
| Avg. position | Your place in the list of companies an answer names, where 1 is named first. |
| Avg. Citation Rank | The average rank of your cited domain among an engine's sources, where lower is better. |
Rates are computed over answers that carry content; blank answers are kept in exports but left out. The formulas are on Methodology and each term is defined in the Glossary.
Is a mention the same as a recommendation?
No, and the report keeps them apart. Mention rate says whether you were named at all. Avg. position says whether you were named first or last in a list. The Sentiment tab reads how each answer describes you, and Accuracy claims checks what it says against your facts. A brand named fifth with a wrong price is mentioned, not recommended.
What comes after the answers?
- Citations: which pages the engines trust for your questions, and whether yours is among them. See Citations.
- Why you lost: for each question a competitor took, what changed in the answers and the pages behind them.
- Fixes: a ranked list of what keeps engines from naming you, each with a drafted fix and its status, then a check that it worked. See Fixes.
- Websites: what AI crawlers can and cannot read on your site. See Crawler access.