Hi, I’m wondering if i could make AI search answer more correctly.
My understanding of AI search is first LLM changes questions in a search bar into graphQL and then returns the corresponding data as an answer.
We need to tune questions so that we can draw the answer we are looking for and is there any example of use of RAG to tune questions more friendly to your own knowledge graph?
I don't know technical detail but in my image, designing to sandwich LLM between RAG and graphQL might help improve accuracy.
Check the
documentation
Ask the
Community
Take a look
at
Academy
Cognite
Status
Page
Contact
Cognite Support

