Question-only embeddings can miss useful RAG answers
A customer email reply chatbot can use RAG by turning old email conversations into stored question-and-answer pairs. extracts the main questions and answers from each email thread, rewrites them in a more general form, and Voyage creates for them. Near-duplicate pairs are merged when both the question and answer have very high .
For each new email, extracts the core questions, the system searches stored question for similar past question-and-answer pairs, applies , and sends the selected material to the more expensive to draft the reply. The weak point is that small wording changes in a new question can stop the right stored answer from being found. A simpler older setup, where received one large FAQ file, can sometimes work better because all information is always present in the context.
One possible improvement is to store and search for both the question and the answer, not only the question.
Key points
- The system uses old customer email threads as reusable question-and-answer examples.
- handles question , while the more costly writes the final email.
- Searching only question can fail when a question is phrased differently.
- Sending a full FAQ costs more tokens but may avoid missing needed information.
- Adding answer to search is a possible way to improve retrieval.