Is it possible to train an investment model?

I want to train a model about investment, and I want to complete the following functions. Is this possible?
When I enter the stock code, it will automatically dig out the comparison from the financial report and get useful information, such as xx is lowered, xx is raised, and we need to pay attention to this risk. Or frequent changes in the number of top ten shareholders.

thanks any help

Sounds more applicable to Langchain. LLMs from my understanding are prone to hallucinate on quantitative info.


This would be too risky, and also GPT is not good with numbers

As suggested by @charlieevert , Langchain might be a better option