The questions
Every scan asks the same set of questions a buyer in your category would ask when looking for a vendor: questions about the category, questions that start from a problem, head-to-head questions about the competitors you entered, and questions about you by name. Companies in the same category get the same questions, so their results can be compared.
The AI assistants
Each question goes to ChatGPT, Gemini, Claude and Google’s AI Overviews, and we read what a person would see in each one, including the sources it cites. AI gives a different answer each time, so every question is asked three times on every assistant. Your first read comes from the first run; the final read uses all three.
What we measure
- Named in answers: how often AI names you when a buyer asks about your category without mentioning you.
- Share of AI voice: your share of all the times AI names you or the competitors you entered.
- Your site cited: how often your own website is one of the sources AI cites.
- How you are described: each mention is labelled, from recommended strongly to mischaracterized, and every label shows the exact words behind it.
Every figure is counted from the answers themselves. No AI model makes up a number in your report.
The tiers
- Leader: AI names you first.
- Contender: AI recommends you alongside the leaders.
- Overlooked: AI knows you, but rarely recommends you.
- Invisible: AI doesn't name you in answers about your category.
What we never do
- Change a result. Each read is stored once and never rewritten.
- Let a model compute a figure.
- Publish your result. The aggregate study only ever shows totals; see our privacy notice.

