Quick answer: Embedding text into vectors + similarity search — foundation of modern RAG.
Vector search transforms text and data into numerical representations called embeddings, enabling computers to understand semantic meaning rather than just keyword matching. Instead of finding exact text matches, vector search measures similarity between ideas—asking 'how close is this to what I'm looking for?' This forms the foundation of Retrieval-Augmented Generation (RAG), where systems retrieve relevant documents from a knowledge base to answer questions accurately. Vector search powers modern AI applications: semantic search in documentation, recommendation engines that suggest similar products, and intelligent chatbots that understand context. You embed documents once, then perform lightning-fast similarity searches across millions of records using techniques like cosine similarity or HNSW indexing.