This document presents an analysis of 90 days of conversational data from ExcelR to extract insights that can help improve the business. The objectives are to perform topic mining and exploratory analysis to help optimize resource allocation, modify content, and improve services. The analysis includes cleaning the data, performing exploratory data analysis, applying machine learning techniques, and providing insights. Various variables like timestamp, unread status, visitor email, country, age, and chat text are analyzed. Topic modeling using LDA identifies 4 topics: course inquiry, career transformation, assistance, and e-learning/discounts. Classification models like Naive Bayes, Logistic Regression, and Catboost are applied to the chat data tagged with time durations, achieving accuracies above 97%. The