How to read your AI visibility score
Updated 2026-10-27
Your AI visibility score is the share of verified facts about your business that a model returned correctly. A low score for one model usually means missing or inconsistent source content, not that the model is broken. Fix the high-impact gaps first, then re-run the report to confirm they closed.
The match states
- Match: the model returned the fact as it appears on your site.
- Partial: some items matched, others were missing (for example 2 of 3 services).
- Missing: the model did not mention the fact at all.
- Unsupported: the model stated something your site does not support. This can be an outdated claim or an invented one.
Why scores differ between models
ChatGPT, Claude and Gemini draw on different training data and different search results, so one can know you well while another knows little. A big spread between models points to uneven coverage of your business across the web, not just on your site.
What to fix first
- High-impact gaps: core services, locations and identity facts.
- Facts missing from all models: usually absent or buried on your own site.
- Unsupported claims: check whether an old page, listing or directory still states them.
- Facts only some models miss: strengthen third-party mentions and structured data.
Treat it as a trend, not a verdict
Scores are a snapshot and can shift as models update. Re-run after changes and compare. See a worked example in the sample report at /sample-report.