Comprehensive glossary of AI and commerce terminology to help you navigate the evolving language of artificial intelligence.
Glossary Explorer
Browse all glossary terms
Vector Database
A vector database is a specialized data store designed to efficiently index and query high-dimensional vector embeddings — numerical representations of data such as text, images, or audio that capture semantic meaning. Unlike traditional databases that retrieve records by exact field matches, vector databases perform approximate nearest-neighbor (ANN) search, finding records whose vectors are most similar to a query vector in high-dimensional space. Vector databases are the storage backbone of RAG (Retrieval-Augmented Generation) systems and semantic search applications. In commerce, they enable capabilities such as finding products semantically similar to a natural language query, identifying customers with similar purchase histories, clustering support tickets by topic, or retrieving the most relevant policy documents for an AI assistant. Popular vector database options include Pinecone, Weaviate, Qdrant, Chroma, and pgvector (a Postgres extension). Selection criteria include query latency, scalability, filtering capabilities, and integration with the model stack.
Vector Database (Vector Store)
A vector store (also called a vector database) is a data management system purpose-built for storing, indexing, and searching vector embeddings — dense numerical arrays that represent the semantic content of text, images, audio, or other data. Vector stores use approximate nearest-neighbor algorithms (such as HNSW or IVF) to retrieve the most semantically similar records to a query embedding with low latency, even across millions of vectors. Vector stores are a critical infrastructure component in enterprise AI architectures, particularly for knowledge retrieval and personalization systems. In commerce, they serve as the retrieval layer in RAG pipelines (storing product, policy, or knowledge base embeddings), the similarity engine for visual search, and the similarity backbone for recommendation systems. Choosing the right vector store involves tradeoffs between query speed, recall accuracy, metadata filtering capability, scalability, and operational complexity. Many enterprises begin with lightweight embedded options (Chroma, FAISS) and migrate to managed services (Pinecone, Weaviate Cloud) as production scale demands grow.
Vector Embeddings
Vector embeddings are dense numerical representations of data — text, images, audio, user behavior, or other inputs — generated by trained neural networks such that semantically similar inputs produce numerically similar vectors in a high-dimensional space. The distance between two embedding vectors (measured by cosine similarity or Euclidean distance) reflects semantic relatedness: the embeddings for "running shoes" and "athletic footwear" will be close together, even though they share no literal words. Embeddings are the enabling technology behind semantic search, recommendation, and retrieval in modern commerce AI. By converting product titles, customer queries, review text, and behavioral data into a shared embedding space, systems can perform semantic matching that far surpasses traditional keyword search. Embedding models from providers like OpenAI, Cohere, and open-source projects (Sentence Transformers, BGE) can be fine-tuned on domain-specific data to improve relevance for particular product categories or customer vocabularies. The quality of embeddings is a fundamental determinant of downstream retrieval system performance.
Vector Store
A vector store is a data storage and retrieval system optimized for managing vector embeddings — high-dimensional numerical arrays that represent the semantic content of text, images, or other data. Vector stores index these embeddings using algorithms designed for approximate nearest-neighbor (ANN) search, enabling fast retrieval of semantically similar items without exhaustive comparison of all stored vectors. In enterprise AI systems, vector stores function as the memory layer for knowledge retrieval. They are commonly used in RAG architectures to store embedded chunks of documentation, product data, or domain knowledge that an LLM can query at inference time. For commerce platforms, this translates to capabilities such as semantic product search, contextual customer service responses grounded in policy documents, and similarity-based recommendations. The operational considerations for vector stores — including index management, embedding versioning, and update latency — are important architectural concerns as these systems move from prototypes to production scale.
Voice Assistant
A voice assistant is a conversational AI system that accepts spoken natural language as input, processes the speech through automatic speech recognition (ASR) to produce a text transcript, applies natural language understanding (NLU) to interpret intent and extract entities, executes the corresponding action or retrieves the relevant information, and delivers a spoken response generated through text-to-speech (TTS) synthesis. Modern voice assistants—including Amazon Alexa, Google Assistant, Apple Siri, and enterprise-grade conversational platforms—integrate ASR, NLU, dialogue management, backend API calls, and TTS into a low-latency end-to-end pipeline designed for hands-free, screenless interaction. In commerce, voice assistants are both a customer-facing channel and an internal productivity tool with significant untapped potential. Consumer voice assistants handle product reorders, order status queries, and shopping list management, though complex purchase decisions remain challenging due to the lack of visual browsing. Enterprise deployments are increasingly valuable: warehouse associates use voice-directed picking systems to receive pick instructions and confirm actions hands-free, improving speed and accuracy; field service technicians query knowledge bases and log case notes by voice; and contact center agents receive real-time AI-suggested responses through voice-activated copilots. As automatic speech recognition quality improves across accents, noisy environments, and technical vocabulary, and as large language models enable more coherent multi-turn dialogue, voice assistants are evolving from command-response interfaces into genuinely conversational agents capable of handling complex, contextual interactions at enterprise scale.