Artificial intelligence-powered chatbots have become indispensable tools for customer interaction, especially in sectors like insurance. This research introduces a framework that extracts, processes, and leverages policy data to optimize chatbot performance. By harnessing advanced NLP, OCR, and knowledge representation techniques, the system automates the extraction of crucial details like coverage, exclusions, and terms. This extracted data is organized into a dynamic knowledge base, empowering the chatbot to deliver contextually relevant and personalized responses. Experimental findings demonstrate, with up to 94% response accuracy, an 88% cosine similarity, and a 91% F1 score in overall user satisfaction, providing clear evidence of the framework’s effectiveness. This study underscores the potential of integrating intelligent data extraction with chatbot systems, opening doors to innovative applications in insurance and other data-driven fields.