Vector search finds information based on meaning rather than keywords. It converts both the query and stored documents into numerical embeddings and returns the stored content most similar to the query in vector space.
A query about "client payment delay" can surface documents about "accounts receivable aging" or "overdue invoice management" because those concepts are close in vector space even without shared keywords. This is what makes AI search genuinely useful for enterprise knowledge — users can ask questions naturally without needing to know the exact terminology used in documents they have never read.
