Sentiment analysis has become a fast-growing, cross-cutting discipline within the field of artificial intelligence, focusing on the interpretation and classification of emotional responses. In a digitally interconnected world, this field plays an important role in improving human–machine interaction. From customer support to virtual personal assistants, particularly in the smart home sector, it enables devices to respond more sensitively and appropriately to users’ moods and preferences. The study focuses on evaluating the performance of two leading sentiment analysis models, VADER and TextBlob, compared to manual labeling in the smart home domain. The results indicate that the VADER model outperformed TextBlob in identifying positive interactions, while TextBlob maintained consistent performance on both positive and negative sentiment, albeit with a slight bias toward classifying comments as negative. We consider this work a starting point in the exploration of sentiment analysis within the smart home context. The model participation of VADER in future solutions could strengthen trust and elevate user satisfaction, optimizing their interaction and opening the door to an era of intuitive and adaptive home assistance.