The topic modelling analysis here represented is based on texts extracted from the last 5 years of visitors’ reviews on TripAdvisor. The reviews were grouped by destination and refer to each user’s experience with cultural attractions such as museums, buildings, sights, natural attractions, city-guided tours, and shopping and transportation facilities. For this specific analysis, only reviews written in English were selected to represent both foreign and domestic visitors. The output of this Machine Learning (ML) algorithm based on language analysis represents clusters of words that reviewers used together. These groups of words, created not only using frequency but also Artificial Intelligent methods, can suggest to tourism stakeholders topics and information to better address destination promotion and/or to be used as input to create new services and products.