Vector Database
A vector database stores embeddings and retrieves the nearest matches quickly: the infrastructure behind semantic search and RAG at scale.
Designers do not pick vendors, but should understand that freshness, permissions, and chunk boundaries in the index directly affect answer quality.
Share
What it means
Specialized storage optimized for similarity search over millions of vectors, often with metadata filters (user, folder, date).
Why designers should care
Design for index lag (“Docs updated 5 min ago may not appear”), access control (users only retrieve what they can read), and multi-source fusion in the UI.
Example
A copilot searches Slack, Notion, and Google Drive separately, then shows three labeled source groups so users know which system supplied each citation.
Common mistakes
- • Assuming all connected sources are searchable immediately after connect.
- • Retrieval across permissions users should not see. Fixed in backend, but surfaced as confusing citations.
- • One blended result list with no source system or date metadata.