Hi,
I’m trying to develop an Assistant API that delivers insights and recommendations based on localized data exclusively provided in tabular formats. Users will interact with the Assistant to inquire about trends, and the assistant will analyze this tabular data to provide responses, for example, “There has been an increase in activity by xx% since last year and …,” accompanied by relevant recommendations.
I am unsure how to best optimize this, it seems there are File Search, Code Interpreter, and Embeddings tools. I would appreciate your advice on the following:
-
Tool Efficiency: Which of File Search, Code Interpreter, and embedding is most effective for interpreting and responding to queries based on tabular data? How should they be integrated for optimal performance?
-
Data Interpretation Enhancement: How can I leverage these tools to enhance the assistant’s ability to accurately and efficiently understand and analyze localized, tabular data?
Thank you.