Semantic Search

Semantic Search

Semantic Search

Semantic search compares the meaning of a query to the meaning of indexed passages using embeddings, returning results that are conceptually related regardless of shared vocabulary.

It is commonly paired with keyword search in a hybrid approach, since exact matching remains superior for names, Bates numbers, and defined terms.

Alternative Names:

Meaning-Based Search, Vector Search

Why it Matters?

Legal records are full of synonym and paraphrase problems that keyword search handles poorly. Clinical terminology varies by provider, deposition testimony describes concepts without using their legal names, and productions use inconsistent vocabulary across custodians. Semantic retrieval closes those gaps, while hybrid search preserves the precision that exact matching provides for identifiers.

Frequently Confused with

Related terms

Frequently asked questions

Is semantic search better than keyword search?

Is semantic search better than keyword search?

Neither is uniformly better. Semantic search handles paraphrase and synonym; keyword search is more precise for names, numbers, and exact phrases. Hybrid approaches use both.

Does semantic search satisfy e-discovery search obligations?

Does semantic search satisfy e-discovery search obligations?

It can be part of a defensible process, but adequacy depends on validation through sampling rather than on the retrieval method chosen.