Merchants care about recommendations and citations
In AI search, teams want to know whether ChatGPT, Perplexity, Gemini, or Doubao recommends their products and URLs to real shoppers, not only how many bot requests appeared.
- Recommendation exposure is closer to a new acquisition position.
- Crawler collection is closer to asset safety and server-load monitoring.
- AI referrals are downstream traffic outcomes that need order, product, and GSC context.
Evidence levels keep the dashboard honest
AI visibility is affected by cache, platform logs, bot behavior, and model browsing capabilities. One blended number creates false certainty.
- Confirmed needs URL, source, or platform evidence.
- Likely signals can guide review but should not become hard exposure.
- Weak samples are useful for review, not commercial claims.