What is a vector database?
A vector database stores embeddings and finds the nearest ones to a question. It is the index behind 'search our docs by meaning.'
Also called: vector db
In plain English
Your normal database stores rows: patient id, time, status. A vector database stores the numeric meaning of paragraphs so an assistant can pull the right paragraph. You may never buy one directly. Notion, a help desk, or a RAG tool may include the index. You do not need a new database because a blog said agents require one. If the shop's truth is a sheet and a folder, start there. A vector store helps when you have more passages than a person will skim. It does not replace the price list, and it goes stale on the same day the price list does. Plan the update, or do not add the store.
How an owner would use it
Ask the vendor where the index lives and how you delete a file from it. If you remove an old fee schedule from the folder but it stays in the index, the assistant will keep quoting it.
Common mistakes
Buying a separate vector database for a ten-page SOP set is a project, not a win. Most owners already have the index inside a help desk, Notion, or the chat product. The failure mode is deletion: you remove a file from the folder and the index keeps it. Ask how to delete and how long that takes. If the vendor cannot say, do not put prices in that index.
A worked example
An optometry office
The old lens price
This is an illustration, not a customer result. The office replaced a fee PDF in the folder. The assistant still quoted last year's progressive-lens price because the index had not been refreshed. They removed the old file from the knowledge tool and asked the question again before anyone repeated the number.
Related terms
Prompts and guides
Questions owners ask
What is a vector database?
It is a store built to search embeddings, so an app can retrieve text that is close in meaning to a question.
How do I use a vector database?
Most owners use the one inside their document tool. The practical job is keeping the source files current and removing files you do not want quoted.
Is a vector database the same as my spreadsheet?
No. A spreadsheet stores the rows you type, like a price and a SKU. A vector database stores numeric meanings of passages so a question can find a paragraph. Your prices should still have a home in the sheet or the SOP.
Product guides are on AI apps for business.