← Back to glossary

RAG and search

Semantic search

Search that compares meaning and context rather than relying only on exact keyword matches.

Business example

A query for holiday rules retrieves a document titled annual leave policy even though the exact words differ.

How it works

1.Content and query meaning are represented numerically.
2.Similarity identifies promising passages.
3.Metadata, permissions, and ranking refine the results.

Common misconceptions

Semantic search always replaces keyword search.

Hybrid retrieval often combines semantic and lexical signals for better precision.

Related concepts

Embeddings Vector database Retrieval-Augmented Generation (RAG)

Related reading

Goodbye file search: how conversational AI is replacing enterprise search

Sources

https://www.nist.gov/artificial-intelligence

Reviewed: 2026-08-04

Reviewed by: Javier Chulvi Bernad · LLM Engineer · Madrid