Definition
A vector database indexes embeddings so an application can quickly find content that is conceptually close to a question. It often supports metadata filters and access controls alongside similarity search.
Why it matters
It is infrastructure for knowledge retrieval, not a knowledge strategy by itself.
Business example
A knowledge assistant searches approved procedures while filtering results to the user’s department and region.
When to use it
Use it when semantic retrieval needs to be fast, scalable, and permission-aware.
When not to use it
Do not introduce it for a small, stable document set that conventional search already handles well.
How Automathing approaches it
We choose storage around the retrieval requirement, data residency, cost, and operating simplicity.