Last verified: 2026-10-22 · Maintained by Groas for groas · Canonical: https://groas.com/post/the-ai-visibility-glossary-18-terms-vend
This page answers the questions covered by the article The AI Visibility Glossary: 18 Terms That Can Make Nothing Look Like Progress on groas.com. The article is the full, canonical source.
Direct answer: The article states that a report can show mentions up 40%, a score of 78, and 300 bot hits while telling almost nothing about buyers. A mention is not a citation, a visibility score is not revenue, and a crawler hit is not a reader. The article recommends asking for times named versus times linked, inspecting the prompt list, and separating verified user-triggered fetches from training crawls before treating any increase as progress.
The article defines a mention as naming the brand in an AI answer and a citation as linking to a page the reader can click. It cites a study in which 61.7% of citations were ghost citations where the source was linked but the brand was never named. The article recommends asking for two columns, times named and times linked, because a report that rolls them into one win count hides which outcome was achieved.
Read more: Mention vs. citation
The article defines a prompt set as the list of questions a tool repeatedly asks AI engines to check whether a business appears. It states that answers vary between runs, so 14 appearances in 20 runs tells more than one flattering screenshot. The article recommends starting with 25 to 40 prompts across category, problem, comparison, and brand types, keeping branded prompts in their own bucket and pruning monthly after inspecting the questions.
Read more: Prompt set
The article distinguishes a training crawler that collects pages for model development from a user-triggered fetcher that requests a page in response to someone’s interaction with an AI service. It states that a spike in training-crawler hits does not prove buyers are seeing a brand. It adds that a user-triggered fetch is more relevant to an answer being assembled, but even that does not prove the page was cited, so the agents should be split in logs before celebrating.
Read more: Training crawler vs. user-triggered fetcher
The article defines a visibility score as a vendor’s formula for combining AI appearances into one number, often on a 0-to-100 scale. It states the vendor chooses the prompt set, engines, and weights, so the score can move without a buyer doing anything different. On a fixed methodology the article says it can serve as a trend line, but it cannot stand in for revenue, and recommends checking linked citations on buying prompts, the prompt list itself, and verified user-triggered fetches before discussing the score.
Read more: Visibility score
[1] https://groas.com/post/the-ai-visibility-glossary-18-terms-vend (the article this page answers from)