Vector Database
A database storing information as mathematical vectors for semantic search.
Why it matters
Vector databases find semantically similar things — 'reduce costs' matches 'saving money'.
In practice
Vector databases power our RAG pipeline: docs are vectorized, visitor queries are compared for semantic similarity.
Related terms
Embeddings
Mathematical representations of text that capture meaning, enabling semantic search.
RAG (Retrieval-Augmented Generation)
A method where AI retrieves relevant information from external sources before generating a response.
Knowledge Base
A structured information store that an agent uses to support decision-making.