TL;DR. Once you have a flagged prompt, look at its 8 signals (Detection, Position, Sentiment, Mentions, Citations, Stability, Engine agreement, Competitor dominance) and map the pattern to one of 10 diagnostic rules. That tells you which gap to fix and which module to go to next.
The 8 signals to read
For your priority prompt, record one status per signal using the live KPI strip Rankscale shows for each prompt in the main dashboard.
| Signal | Options |
|---|---|
| Detection Rate | Yes (all engines) / Partial / No (not found) |
| Position | #1–3 strong / #4–6 mid / #7+ weak / Not Found |
| Sentiment | Positive / Neutral / Mixed / Negative |
| Mentions (out of 7 runs) | 6–7 / 3–5 / 1–2 / 0 |
| Citations (out of 7 runs) | 6–7 / 3–5 / 1–2 / 0 |
| Stability | Stable / Volatile / Too early |
| Engine agreement | Consistent / Mixed / Conflicting |
| Competitor dominance | None / One competitor wins / Multiple win |
The 10 diagnostic rules
Compare your signals against the rules. The first match is the diagnosis.
- Poor detection on a branded prompt. AI cannot find your brand when you ask about your own brand name. Diagnosis: entity signal problem. Engines do not associate your brand with itself. Go to Module 6 (Entity gap).
- Poor detection on a non-branded prompt. AI does not bring your brand up when the topic is your category. Diagnosis: topic association gap. Go to Module 5 (Content gaps) to check whether the content even exists.
- No mentions and no citations. You are not in the answer and your pages are not in the source list. Diagnosis: full invisibility on this topic. Branded: crawler or entity problem. Non-branded: content-existence problem.
- Mentions missing, citations present. Your pages are cited but your brand is not mentioned on this topic. Diagnosis: your content gets cited but lacks your brand name to get mentioned. Check topic-specific content and schema scope.
- Mentions present, citations missing. AI mentions you in the answer but will not cite your pages as a source. Diagnosis: Trust Gap. Your content is not citable enough. Go to Module 5 (BLUF + Justification) in the Content module.
- Weak position. You are mentioned but ranked #7+ or "Not Found" on the key prompts. Diagnosis: a competitor has content AI prefers. Mimic and make the top URLs (Citations Tab) even more optimised to AI search engines (Module 5).
- Volatile position across runs. Your rank fluctuates every time the prompt is re-run. Diagnosis: source-pool instability. Third-party coverage is thin. Go to Module 6 (Citation gap).
- Negative sentiment. AI describes your brand negatively. Diagnosis: reputation problem. Go to Module 6 (Sentiment gap).
- Engines disagree on sentiment. One engine describes you positively, another negatively. Diagnosis: source conflict. The lowest-scoring engine is pulling from a worse source pool. Check which domains it cites.
- A competitor dominates. A named competitor consistently wins the prompt across runs. Diagnosis: format gap. Their page is more extractable. Reverse-engineer it in Module 5 (content) or Module 4 (page/technical) as needed.
If no single pattern triggers
Multiple weak signals, no red flags: several minor gaps. Prioritize content freshness and entity consistency. No red flags at all: minor gaps only; optimize existing content.
Do this now:
Pick your highest-priority flagged prompt, record the 8 signals, match the pattern. Write down: "Diagnosis = [rule number and name]. Next module = [4 / 5 / 6]." That is one line of documentation per flagged prompt.
Start improving your AI visibility today with Rankscale.
Get started