A developer builds a RAG (Retrieval-Augmented Generation) pipeline for answering security policy questions. Which component is responsible for converting both the user query and policy documents into a common numerical representation for similarity search?

Why A is correct

In a RAG pipeline, the embedding model converts both the user query and document chunks into dense vector representations (embeddings) in a shared semantic space. The vector store persists these document embeddings and enables fast approximate nearest-neighbor search. The reranker (optional) re-scores retrieved candidates for relevance. The LLM generator receives retrieved context plus the query and generates the final answer. The embedding model is the foundational component that makes semantic similarity search possible.