AI-Powered Chatbot for Automating Responses to Banking Regulations
Degree
Master of Science in Data Science
Department
Department of Computer Science
Faculty/ School
School of Mathematics and Computer Science (SMCS)
Date of Submission
Fall 2024
Supervisor
Khawaja Abdul Hafeez, Visiting Faculty, Department of Computer Science
Abstract
Banking regulatory compliance is a highly sensitive yet very complex area of activity, which calls for the exploration of numerous directives issued by the State Bank of Pakistan. Compliance-related queries in the traditional manner are both time-consuming and inefficient. This project introduces an AI-powered chatbot that simplifies the process of retrieving regulatory information by utilizing OCR, OpenAI's GPT-4, Pinecone, Cohere, and Tavily. The chatbot extracts text from SBP documents, generates question variations, indexes summaries in Pinecone, retrieves relevant information from MongoDB, and ranks documents for the best responses. If internal documents do not contain answers, Tavily fetches information from external sources. This RAG-based pipeline ensures that the responses are accurate and quick, thereby making regulatory compliance efficient.
Keywords: Automated summarization, web scraping, OCR, RAG-based LLM, Pinecone, Tavily, chatbot, data-driven insights, Mongo DB
Document Type
Restricted Access
Submission Type
Research Project
Recommended Citation
Zaidi, S. (2024). AI-Powered Chatbot for Automating Responses to Banking Regulations (Unpublished graduate research project). Institute of Business Administration, Pakistan. Retrieved from https://ir.iba.edu.pk/research-projects-msds/40
The full text of this document is only accessible to authorized users.
