← Glossary/Architecture

Defined term

Semantic search

Search by meaning, not by keyword overlap.

Semantic search retrieves results that match the meaning of a query, not just exact terms. It uses embeddings (vector similarity) and often combines with keyword (BM25) for hybrid retrieval. Hybrid retrieval wins most production benchmarks because it captures both lexical precision and semantic recall.

When it matters

When users describe what they want in natural language rather than typing exact keywords. Replaces or augments keyword search where intent matters more than exact terms.

Real example

An internal knowledge portal where 'how do we handle returns for damaged goods in international orders' returns the same top-3 results as 'damaged international return policy' — keyword search misses the second; semantic search hits both.

KPIs to watch

Semantic search precision@5 (>0.8 target), query reformulation rate (target: -40% vs keyword), zero-result rate (<5%).

Related terms

See it in action

We use this every week

Book a 30-min call and we'll walk you through how Semantic search shows up in a real engagement we're running.

Book a 30-min call