Jeff Chow, TripAdvisor vice president of product and consumer experience says machine learning - which powers AI - is used to “derive meaning” out of TripAdvisor’s massive data sets, and AI is then “married to other techniques such as UX, interface design to make it more humanized.” A process called collaborative filtering matches users with specific interests to reviews that include similar content - in essence looking for patterns to surface relevant reviews to each traveler. Sentiment analysis helps determine where users are in the planning cycle as well as what types of reviews they’re looking for: Someone narrowing down a selection of 10 hotels, for example, likely won’t need to read full reviews for each property, however a traveler ready to make a purchase might want more information. Get the full story at Phocuswire