An AI-Powered Sentiment Analysis System to Measure the Effect of ICT Communication on Employee Self-Efficacy

The emergence of ICT-enhanced communication has played a significant role in shaping the mode of behavior, emotional involvement, and self-observed competence within the organizational context. The proposed method is an AI sentiment analysis system that would analyze text-based interactions of emails, chat, and meeting minutes to gauge the effect on employee self-efficacy. The architecture has a combination of BERT to capture the contextual meaning and GRU to track the emotion sequentially. It handles more than 25,000 labeled messages of organizations and derives the sentiment polarity, the confidence scores as well as engagement behavior. The proposed model uses a self-efficacy prediction score of 0.84 and has its accuracy level in sentiment classification standing at 93.4%. Positive sentiment makes up 41.9 percent of all communications with a strong relationship to the higher selfefficacy indices and quicker response to tasks. The system also brings down emotional misclassification relative to past procedures. The technology based on this AI concept can allow organizations to track employee morale in real-time and preventively address emotional states and performance with smart insights of ICT communication.