This document discusses an intelligent auto-response model for categorizing SMS messages into five categories: ham, spam, info, transactions, and one-time passwords, using the multi-layer perceptron (MLP) algorithm. The proposed model addresses limitations of previous binary and multi-label classification approaches by implementing a multi-class categorization that achieved 97% accuracy on a dataset of 7,398 messages. The paper highlights the significance of automated feature extraction and various machine learning techniques used in existing SMS classification models.