International Journal of Scientific Research and Engineering Development

The quick growth of digital documents has made efficient information retrieval a major challenge for both organizations and individuals. Traditional keyword-based search techniques often fail to deliver complete and contextually relevant results. To address this limitation, this study proposes an intelligent system titled AI-Driven Document Question Answering and Content Retrieval Using Large Language Models and Vector Embeddings. The system leverages the capabilities of large language models (LLMs) to enable accurate, contextaware, and interactive exploration of document content. The proposed approach processes unstructured documents by transforming them into semantic representations using embedding techniques and storing them in a vector database for efficient similarity-based retrieval. When a natural language query is submitted, the system identifies the most relevant document segments based on semantic meaning rather than simple keyword matching. The LLM then generates precise and human-like responses based on the retrieved information, ensuring both accuracy and readability. Moreover, the proposed framework enhances search efficiency, reduces manual effort, and improves user experience through a conversational interface. The system is designed to be scalable and supports multiple document formats, making it applicable across various domains such as academic research, enterprise knowledge management, customer support, and technical documentation. Overall, this work demonstrates how AI-driven retrieval systems, powered by large language models, can transform traditional document search into an intelligent knowledge discovery platform.

Sukrutha S Bhat, Prashanth Ankalkoti, "AI-Driven Document Question Answering and Content Retrieval Using Large Language Models and Vector Embeddings" International Journal of Scientific Research and Engineering Development, V9(4): Page(911-913) July-August 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.