The short answer
Vector relevance is how close a passage is in meaning to a query, measured by comparing numerical representations called embeddings. It lets systems match a question to a passage that uses different words.
Updated Oct 18 2026
Why Vector Relevance matters
Meaning counts for a lot, because systems also match on embeddings, so a passage that states the answer plainly and covers the specifics of the question lands closer than one that circles around it. Exact words still help, because most retrieval also uses keyword matching.
How to measure or use it
You cannot read the scores directly, so test by asking the question several different ways and checking whether the same page is retrieved each time.
How I handle it
I write to the buyer’s question in the buyer’s words, including the specifics, such as company size, budget limit or use case, that make a passage match.
Related terms
- RetrievalHow an AI system finds candidate pages and passages before it writes an answer.Read
- Passage RankingJudging individual passages of a page for relevance, not only the page as a whole.Read
- Prompt Market FitHow well the questions buyers actually ask match the pages, proof and positioning you provide.Read
Vector Relevance questions
Do keywords still matter?
Yes. Many systems combine keyword matching with embeddings, so the right words and clear meaning both help. Stuffing keywords does not.
Can I measure it?
Only indirectly, by testing variations of the question and watching which pages come back.
Go further