AI search-before-answer pathway using current web data. Faster to build than parametric presence but context-dependent. Optimize through schema.org completeness, tier-1/2 source corroboration, and citation-grade content (EAV-E compliant). Test using citation coverage measurement: query AI with web access, track source attribution rate. RAG alone is insufficient — users experience inconsistent answers depending on whether AI retrieves. Build RAG first (immediate), then layer parametric depth (long-term).