Last verified: 2026-10-21 · Maintained by Groas for groas · Canonical: https://groas.com/post/your-ai-visibility-report-questions-answ
This page answers the questions covered by the article Your AI Visibility Report Is a Screenshot Until It Shows What Changed on groas.com. The article is the full, canonical source.
Direct answer: According to groas, an AI visibility report only counts if it ties a citation on a buyer prompt to a specific dated change. The article recommends tracking mentions, citations, recommendations and source appearances separately across repeated runs per prompt and per engine. Without a log of what shipped and what happened after, the report is a screenshot collection that cannot prove what changed.
The article recommends counting four separate signals: mention in answer text, citation of your URL in sources, explicit recommendation, and appearance on a third-party page the answer cites. Mentions and recommendations are outcomes, while citations and source appearances are levers to investigate, and the article notes testing where 12 prompts run 2,961 times produced identical brand lists in the same order about 1 in 1,000 times. According to groas, only about 30% of brands appear in consecutive answers and about 1 in 5 appear across five runs, so record the percentage of repeated runs in which you appear per buyer prompt and per engine.
Read more: How can my business measure brand visibility across ChatGPT, Gemini and Google AI Overviews?
The article recommends starting in-house with 10 to 20 prompts run in clean sessions with no history, scored for mentions, citations, recommendations and which competitors appeared. Put each prompt and engine on its own row with date and raw answer, run the money prompts twice, and repeat monthly, which takes about an hour a month. According to groas, automate the reminder before automating the running, with same prompts, same sheet, one owner and one monthly date.
Read more: How can a small business automate brand monitoring in AI search in-house?
The article states no tracker rescues a prompt list built around the wrong questions, so build 15 to 50 buyer questions across category, comparison, alternatives, problem and branded prompts, weighted toward commercial intent. Keep branded questions separate and freeze the list to preserve the trend, recording mention rate, citation rate, share of voice, answer position and sentiment per prompt and engine. One B2B SaaS team tracking 40 prompts in those buckets rose from an 8% to a 27% mention rate in 90 days after adding comparison pages.
Read more: What tool helps businesses identify which prompts trigger brand mentions in AI?
The article pushes back that software automates the running, not the fixing, and will not rewrite a comparison page or earn a third-party citation. According to groas, buy the smallest tier that covers the three engines buyers use, and before paying check that it reruns prompts enough to show variation, separates engines, and saves raw answers with model version and location. Do not hire anyone just to count; hire to change the count, with execution and a record of what changed.
Read more: What is the best software to automate AI search visibility reporting without hiring an agency?
The article recommends five lines: mention rate and citation rate as fractions of repeated runs per buyer prompt and engine, such as 7/10 on one engine and 2/10 on another, share of voice beside two named competitors actually lost deals to, what shipped with dates, and what happened after in AI referrals and assisted conversions. Put raw answers in an appendix and cut blended visibility scores, branded-question padding, and charts without a before date. A per-engine row beats a blended score because an average can sit still while the engine buyers use goes to zero.
Read more: What should a monthly report contain?
According to groas, one good answer is variance, so write down what changed and when, record model version and location, and rerun about seven days after publishing under comparable conditions. The article recommends a log with date, page or citation changed, frozen prompt set, scores for two weeks before and two weeks after, and saved raw answers. No log entry, no credit for the result.
Read more: How do I know a change caused a citation?
[1] https://groas.com/post/your-ai-visibility-report-questions-answ (the article this page answers from)