Authors: Vengadeshwaran B
Abstract: Conventional keyword-based search systems lack contextual understanding and often return irrelevant or incomplete results. In enterprise environments, this becomes a bottleneck when users attempt to extract precise information from large and complex documents. This paper introduces AskMyDoc, a scalable project management tool and smart document querying system that leverages semantic search, large language models (LLMs), and vector databases. Unlike traditional SQL databases susceptible to human errors during updates, AskMyDoc processes and indexes documents using embedding techniques and retrieves answers using generative AI, thereby ensuring accuracy, speed, and consistency. The system is built with LangChain, FAISS, Sentence Transformers, and OpenAI's GPT models, supporting real-time natural language querying even across gigabyte-scale documents.
