How to handle extremely large extracted document data in an agentic LLM system?

artificial-intelligence large-language-model agent information-retrieval retrieval-augmented-generation