Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{178847,
        author = {P.M.ASWATHI SWARNA SREE and AKSHAYA J and ANITHA GRACE S},
        title = {A BANKING CHATBOT USING RETRIEVAL AUGMENTED GENERATION},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {11},
        number = {12},
        pages = {5104-5110},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=178847},
        abstract = {This paper presents a streamlined architecture for a multilingual banking chatbot that relies on text-only input and supports optional Text-to-Speech (TTS) output, removing the need for Automatic Speech Recognition (ASR). The design leverages a Retrieval-Augmented Generation (RAG) framework combined with traditional and semantic retrieval techniques for robust multilingual understanding and contextual response generation. The system uses Langdetect for language identification, FastText for intent detection and named entity recognition (NER), BM25 for keyword-based search, and LanceDB for vector-based semantic retrieval. A hybrid retriever merges results from both retrieval methods, which are then processed by the Mistral Large Language Model (LLM) within the RAG pipeline. Optional voice responses are synthesized using Kokoro TTS. This simplified yet powerful design reduces latency and complexity, while maintaining accessibility and compliance in secure banking environments.},
        keywords = {Multilingual chatbot, Text-to-Speech (TTS), Retrieval-Augmented Generation (RAG), BM25, LanceDB, FastText, Mistral LLM, hybrid retrieval, banking automation, conversational AI.},
        month = {May},
        }