TL;DR. Five tabs, five questions. Visibility = how am I doing overall? Mentions = is AI describing me? Citations = is AI sourcing my pages? Sentiment = how is AI describing me? Position = where am I in the answer? Start every weekly review at Visibility, drill into whichever tab moved most.
The five tabs
| Tab | The question it answers | What "good" looks like |
|---|---|---|
| Visibility | Overall health per prompt group | Above Standard for your category depth (lesson 2) |
| Mentions | Is my brand named in answers? | Your name appears ≥ top 3 competitors on non-branded, ≥95% on branded |
| Citations | Are my URLs sourced? | Your domains sit in the top 5 cited sources on your prompt group |
| Sentiment | What tone does AI use about me? | ≥65% = positive. 65%>x>55% = neutral. Below that = reputation risk |
| Position | Where do I appear in the answer? | #1–3 position on branded, ≥#6 on non-branded niche |
The weekly 5-minute read
- Open Visibility. Scan for any prompt that flipped RED this week.
- Check Detection Rate on branded prompts. Anything <90% is an entity emergency, fix first.
- Open Mentions. Who moved up or down in Share of Voice?
- Open Citations. What new domain is being cited this week? That is a new PR target.
- Open Sentiment. Any recurring negative keyword >3 times in 7 days? Source-trace it.
Five minutes, one page of notes. That is enough to run the week.
What to ignore
- Day-to-day fluctuation inside the same status tier. Visibility scores are probabilistic. A single-digit shift without a status change is noise.
- Perfect coverage across all engines on day 1. Detection takes 2–4 weeks to settle once tracking starts.
- Metrics on a prompt group that is not a business priority. If you do not plan to ship fixes for it, do not read it. Just continue to collect data for it.
Do this now:
Open your dashboard. Run the 5-minute weekly read above.
Start improving your AI visibility today with Rankscale.
Get started