Definition
A vector embedding represents content as a set of numbers that describe its characteristics and meaning. This allows a system to compare the similarity of text passages, documents, products, or other objects even when they do not use exactly the same words. Embeddings are often used in semantic search and retrieval-augmented generation (RAG) solutions that need to find material related to a user’s question. Result quality depends on the model, the way content is divided, the freshness of the index, and access filters. A vector representation does not replace access controls or human verification of an AI-generated answer. Explore AI solutions.