Overview
Competitor rows are built from the same captured responses that power your own score. Any domain that appears alongside your brand in at least three prompts is promoted to a tracked competitor and given its own detail page.
Extraction rules
Competitor detection is deterministic and runs on response text plus citations. The parser filters common noise domains (wikipedia, reddit, social networks) so only commercial peers surface.
- Threshold
- Minimum three distinct prompt appearances.
- Scope
- Response body + citations, not prompts.
- Normalization
- Root domain only; subdomains are merged.
Competitor list view
The list view ranks competitors by share of voice — the percentage of prompts where they appear before your brand. Columns are sortable, and the list exports to CSV from the Data Controls panel.
Competitor detail evidence
Each competitor row opens a detail page showing every prompt where they out-ranked you, the response excerpt, and the position delta. Use it to understand why specific prompts are going their way.