Kondreddygari Archana 2 1 (January - March) In todays world humans present their opinions in social media every day a large amount of text data is dumped these user text is used to Train a CBT model to classify emotions Cognitive Behavioural Therapy CBT is a widely applied proofbased practice to address human mental health problems using structured interventions CBT incorporates advanced AI and deep learning techniques to enrich the CBT process For classifying usergenerated text into specific emotional categories based on analysing user intent the CBT Model has utilized NLU models like BERT RoBERTa It uses datasets from Kaggle Reddit for emotion detection The CBT model now generate response using XLNet based on user intent and provide a positive response to lead a healthy life while existing model only classify where the user input text is depressed or Not Depressed but proposed model not only classify user input but also provide suggestions to make users mental health stable This intervention helps in considering AI is more comfortable and provide insight suggestion compared to psychotherapists Copyright (c) World Journal of Pharmaceutical Seiences. All rights reserved 2026