Designing and evaluating an ai "question and answer" system

Hi there,

I want to design an AI customer answering system. I want to feed the history of previous questions and answers to this system.
What would be your choice in choosing the best system between these options:
– LangChain
– LlamaIndex
– LangGraph
And please help me with choosing the best evaluation system between these options:
– Trulens

Thanks in advance.

LlamaIndex and Trulens would be a good combination!