Retrieval inaccuracy

I specifically have a vector with ID: aaa24e8f-142c-49af-9789-59888c408ce0 with the below metadata

but when I query by bot for “give me an AI tool for Mental wellness” it returns a response as below.

Currently, it seems we don’t have a specific AI tool for mental wellness in our database.

I am querying the correct index and namespace.

why is the correct vector not being retrieved from my Pinecone serverless index?

I saw the trace on Langsmith and I find the correct namespace and index being retrieved but the results don’t
make sense based on similarity search (which is what I have set to). Pls refer screenshot (i was able to upload only one screenshot). thanks

My index name is “elevatespot-1”
namespace: “ai-tools”
namespace: “elvspot-abt-blog-artic”
I tried querying with both the namespaces (both have the tool name and function) (the “ai-tools” namespace has a detailed info and “elvspot-abt-blog-artic” has only tool name, category and function)

Hello, I figured the reason for the error. my index was embedded using OpenAI text-embedding-3-small and my user query was embedded using text-embedding-ada-002. I corrected the embeddings for user query to use text-embedding-3-small and now the retrieval works fine.

Thank you for posting your solution!