Reciprocal Rank Fusion (RRF)
The fusion algorithm that merges keyword and semantic search rankings into one consolidated results list.
Plain Language Definition
RRF is the mathematical formula that merges a keyword search list and an AI semantic search list into one combined ranking. Rather than picking a winner, it gives credit to documents that rank well in multiple retrieval methods. When an AI answer engine uses hybrid retrieval, RRF is typically the algorithm that decides the final order of candidate documents.
Technical Definition
Consensus ranking algorithm that merges multiple independently-ranked candidate lists by summing reciprocal rank scores (1/(k+rank)) per document across all lists — rewarding consistent high placement over strong-but-narrow performance in a single system.
Why This Matters for AI Search Visibility
RRF rewards content that performs well on both keyword precision and semantic relevance simultaneously. Content that is terminologically precise and semantically rich will consistently surface in the merged RRF ranking that feeds AI answer generation.