Vector search is a retrieval method that finds documents semantically similar to a query by comparing numerical vector representations (embeddings) rather than exact keyword matches, enabling natural language queries to surface conceptually relevant results.
Last reviewed: 2026/05/19
An embedding is a numerical vector representation of text — such as a word, sentence, or document — produced by a machine learning model, enabling AI systems to measure semantic similarity between texts and retrieve relevant information.
Legal PracticeLegal research AI is software that uses artificial intelligence to help lawyers find, analyze, and synthesize legal authority — case law, statutes, regulations, and secondary sources — with greater speed and comprehensiveness than traditional keyword search.
Tech / ModelRetrieval-Augmented Generation (RAG) is an AI architecture that combines a retrieval system — which fetches relevant documents from a specified corpus — with a generative language model that produces answers grounded in those retrieved documents, rather than relying solely on the model's training data.
AI-powered legal research with citation-validated answers from Westlaw.
Conversational legal research with real-time Shepard's citation validation.
Cloud eDiscovery with AI predictive coding and document summarization.
Thomson Reuters' GPT-backed legal research and drafting with Westlaw integration (relaunched as CoCounsel Legal, 2025).
Move from this definition to role-based legal AI shortlists and the selection criteria that matter for each type of legal team.
Am Law 200 and global firm workflows: accuracy at scale, security compliance, and matter-level auditability.
Last reviewed: 2026/05/19. Definitions are written by the LawyerAI Editorial team. Commercial relationships are disclosed and do not determine editorial scores or conclusions. See our Sponsorship & Affiliate Disclosure.
Vector search is a retrieval method that finds documents semantically similar to a query by comparing numerical vector representations (embeddings) rather than exact keyword matches, enabling natural language queries to surface conceptually relevant results.
Vector search underpins the natural language query capabilities that make modern legal research AI different from traditional Boolean search databases. When a lawyer types a question into a legal research AI — "What standard applies to preliminary injunctions in the Ninth Circuit?" — the system uses vector search to find the most semantically similar cases and statutes in its database, not just those containing those exact words.
This matters for legal research because legal language is not always consistent. Courts and commentators use varied terminology for the same concepts; statutory language may differ from common law language; older cases use different vocabulary than recent ones. Vector search can surface relevant authority that keyword search misses.
In e-discovery contexts, vector search enables concept-based document review: identifying all documents related to a specific topic or event regardless of the precise terminology used. This is particularly valuable in cases where the relevant documents may use informal, coded, or varied language to discuss the key issues.
For lawyers evaluating legal AI tools, the quality of the vector search — meaning how accurately it retrieves the most legally relevant materials for a given query — is one of the most important differentiators between platforms. Poor retrieval quality produces answers that are coherently generated but poorly grounded in the most relevant authority.
Vector search is embedded in the retrieval layer of most modern legal research AI tools. Westlaw Precision AI and Lexis+ AI use vector search to identify relevant cases within their legal content databases, pairing semantic retrieval with their traditional database search capabilities for comprehensive coverage.
E-discovery tools like Relativity AI and Everlaw deploy vector search for conceptual document clustering and relevance ranking — grouping documents discussing similar topics even where no common keyword thread exists.
The performance difference between tools often comes down to the quality of the underlying embedding model and the design of the retrieval pipeline. Tools that combine vector search with traditional keyword search (hybrid search) often outperform pure-vector approaches on legal tasks, because legal research frequently requires both conceptual relevance and exact terminology matching.
Lawyers cannot directly observe the vector search layer but can evaluate its quality through testing: querying for a concept with known authoritative cases and checking whether those cases appear in the results.