Vector Database

What this means in practice

When text, images, or other content is processed by an AI model, it can be converted into a numerical vector that captures its semantic meaning. Vector databases enable fast similarity search -- finding the most relevant documents, products, or answers based on meaning rather than exact keyword matches. They are essential infrastructure for RAG (Retrieval-Augmented Generation) systems, semantic search applications, and recommendation engines. Vector databases are part of the emerging AI infrastructure landscape assessed in COMPEL Domain 10, with organizations at Level 4.5+ typically supporting vector storage for embedding-based retrieval workloads.

Why it matters

Vector databases are essential infrastructure for RAG systems, semantic search, and recommendation engines, enabling fast similarity search based on meaning rather than exact keywords. As generative AI becomes central to enterprise operations, vector database capability determines whether organizations can build grounded, accurate AI applications. Without vector storage, organizations cannot implement the retrieval architectures that modern enterprise AI applications require.

How COMPEL uses it

Vector databases are assessed in COMPEL Domain 10 (Data Infrastructure) within the Technology pillar, with organizations at Level 4.5+ typically supporting embedding-based retrieval workloads. During the Model stage, vector database requirements are evaluated alongside RAG architecture design. The Produce stage deploys vector infrastructure, and the Evaluate stage monitors retrieval performance and relevance quality.

Related Terms

Other glossary terms mentioned in this entry's definition and context.

• Embedding
• Evaluate Stage
• Generative AI
• Model Stage
• Produce Stage
• RAG (Retrieval-Augmented Generation)