For a university project, I need to compare the performance of different APIs (Copilot, Gemini, and ChatGPT) in extracting information from input data and using it to generate answers, summaries, etc.
The input is a .csv file containing a series of transactions, and I want to answer questions such as:
What is the total amount of Credits in the financial report?
What is the description of the last recorded transaction?
Which date appears most frequently in the financial report?
How many transactions are related to maintenance services?
I have already completed this with OpenAI’s ChatGPT and Gemini. Since Copilot doesn’t have a public API, I think I will need to use Azure as a substitute. However, given that Azure OpenAI is essentially the same as OpenAI, this might undermine the purpose of the comparison.
Therefore, I’m looking for an alternative solution to perform these tasks using Microsoft Azure services. Is there a library or service that allows me to extract and process such information from the input data (not necessarily in .csv format, as I can preprocess it) and get answers to the same kind of questions?
I have looked up to Document Intelligence (for the input processing) and QuestionAnswering, but I understood that the latter is more adequate for when you need to extract the answer from the text.
