ChatGPT
Claude
PerplexityUnderstand your report
After a visibility report completes, Ranktune summarizes how AI models talk about your brand compared to competitors across the prompts you tracked.
Overall score
The overall or self score reflects how often your brand is mentioned, cited, or recommended across tracked prompts and models. Higher scores mean stronger presence in AI-generated answers.
Scores are relative to your prompt set and competitors — compare trends over time rather than treating a single number as absolute.
Share of voice
Share of voice is your brand’s share of all tracked mentions across you and your competitors:
your prompt mentions ÷ (your mentions + competitor mentions) × 100
Use the Share of voice tab for a competitive leaderboard and trend chart. High visibility with low share of voice usually means competitors still dominate the same prompts.
AI understanding (entity recognition)
AI understanding answers a different question from visibility: Does the model know who you are and what you do?
| Metric | Question |
|---|---|
| AI visibility | Did models mention or recommend you on tracked prompts? |
| Share of voice | Of all brand mentions in those answers, what share is yours? |
| AI understanding | Can models accurately describe your brand and category? |
Ranktune builds the understanding score from:
- Brand knowledge probes (70%) — Fixed questions such as “What is {brand}?” and “What does {domain} do?”, scored against your site’s identity and topics
- Content entity clarity (30%) — How clearly your pages define brands, products, and concepts for models
Use the AI understanding tab in the report to review probe answers, gaps, and detected entities. A high visibility score with low understanding usually means models name you without a clear product story — or the reverse: they can describe you but rarely recommend you.
Provider breakdown
Provider scores show performance by AI platform (for example ChatGPT, Gemini, Claude, Perplexity, and Google AI Overviews). Use this to see where you are strong or weak:
- A low score on one provider may mean content or citations that model favors are missing.
- Consistent gaps across providers often point to category positioning or source authority issues.
Competitor comparison
The report highlights where competitors appear and you do not on the same prompts. Look for:
- Prompts where rivals are named but you are absent
- Shared prompts where you are mentioned but not recommended
- Citation patterns (who gets linked as a trusted source)
Prioritize prompts with high commercial intent where competitors dominate.
Recommendations
Ranktune generates prioritized recommendations based on actual model responses — not generic SEO checklists. Each item ties back to gaps found in the report (missing mentions, weak citations, competitor positioning).
Treat recommendations as a backlog: tackle high-intent gaps first, then re-run or refresh the report to measure change.
Recommendations are prioritized by estimated score lift and tagged to the metric they move (Visibility, Share of voice, or Understanding). Each item should name a target prompt or competitor when possible, plus concrete Fix steps. Closing those gaps and refreshing the report is what raises scores — having more recommendations does not lower your visibility score.
History and trends
History charts show how scores move across refreshes. Use history to:
- Validate that content or positioning changes improved visibility
- Catch regressions when competitors publish new material
- Report progress to stakeholders on a regular cadence
Fetch history in the product UI or with GET /v1/ai/generate/seo-site-report/history — see AI visibility API.