SRCI-1.0: Source Intelligence
Understand which sources support AI answers, how concentrated the evidence ecosystem is and where cross-model consensus emerges.
How to interpret this document
This content describes technical and methodological behavior that is implemented or explicitly planned in the product. When a control depends on configuration, a provider, a secret, a contract or legal approval, that dependency must remain visible.
The core question
Source Intelligence asks: who is teaching AI about your brand? It does not attempt to map the entire web; it describes the sources actually observed in Grounded responses from the experiment.
Citation coverage
Measures the proportion of Grounded responses where at least one structurally valid source was captured.
Owned and third-party sources
Where possible, the platform separates brand-owned sources from external sources. This helps teams understand dependence on owned channels versus independent validation.
Diversity
A Shannon-based score helps identify whether evidence is spread across many domains or concentrated in a narrow set.
Concentration
An HHI-style measure and top-domain share expose excessive dependence on a small source base.
Influence Score
Combines citation share, model coverage, prompt coverage and valid-citation ratio. It is an internal observational proxy and does not replace Domain Authority, PageRank or an external SEO authority metric.
Cross-model consensus
Jaccard overlap and recurrence across models help distinguish a domain repeatedly present in different answer ecosystems from a source appearing only in one provider.
New source
A source is new when it was not present in historical platform snapshots. This describes new observation, not necessarily recent publication.
Opportunities
Coverage gaps, weak owned-source share, high concentration, low diversity and recurring external domains can inform recommendations. No recommendation guarantees future citation.