In 2023, large-scale AI models will bring a wave of transformation to the world, and how to realize their application in vertical fields has become the focus of various efforts. Earlier this week, Ctrip Group, the Chinese tourism giant, released the first large-scale vertical model of the tourism industry - "Ctrip Asking."
The new AI model's current capabilities mainly include two aspects:
Suppose users are unsure about their travel plans. In that case, the model can recommend travel destinations, hotels, scenic spots, itinerary planning, and real-time discounts based on geography and theme characteristics. If users have specific ideas, the algorithm can offer options to match their preferences.
Another important feature is providing intelligent query results for users with clear needs. Users can use natural language in the form of long sentences in text and voice to search for airline tickets and hotel products under complex conditions.
As the Chairman of the Board of Directors of Ctrip Group, Liang Jianzhang hopes that users can rely on Ctrip to receive trustworthy content and recommendations within the travel industry. Despite the increasing automation of society, tourism remains a demand that cannot be automated, and its significance will only grow as overall social prosperity improves. Ctrip plans to embrace an intelligent society and concentrate on the global tourism industry, using AI to offer competent assistance services before, during, and after travel.
"Ctrip Asking" is a sophisticated model that utilizes 20 billion unstructured tourism data along with Ctrip's existing real-time structural data and historically trained robots and search algorithms. The model was specifically designed for vertical training to ensure high-quality results.
Liang Jianzhang announced that Ctrip invested heavily in addressing the common AIGC (Artificial Intelligence Generated Content) problem and created a dependable answer library for the tourism industry. Using intelligent algorithms, Ctrip manually verified common theme recommendations for hotels, scenic spots, and travel routes to form the "Ctrip Word of Mouth List." As travel is closely linked to hot trends, users also want to avoid price spikes, which led to the creation of the "Ctrip Hot List" and "Ctrip Special Price List." These lists are now representative of content products with unique Ctrip characteristics.
Each list undergoes an average data calculation of 5 million and has been verified by employees in over 30 countries and regions worldwide. This makes it one of the best options for users.
Sun Bo, the Executive Vice President and Chief Marketing Officer of Ctrip Group, emphasized that "Ctrip Asking" includes over 3,000 destinations worldwide, with 45 different themes. These packages offer various options, such as destinations, itineraries, hotels, air tickets, and attractions.
The initial Ctrip hotel word-of-mouth list has achieved impressive statistics. It has been accessed by over 70% of Ctrip users, with 60% of them thoroughly browsing through at least one item on the list. Furthermore, the index has received a recommendation rate of 82%, indicating that many users are willing to share their positive experiences with friends.
According to Sun Bo, Ctrip's travel word-of-mouth list can significantly reduce the time users take to decide about their trip. It has reduced decision-making time from 9 hours to just 6.6 hours, an impressive increase of 27%. The hot list is designed to track changes in travel data and focus on "classic" travel for users. Since its launch, the hot list has become increasingly popular with Generation Z users, with a 1.2 times higher proportion than the classic theme list. Additionally, users born after 2005 have been using the hot list at a rate 1.8 times higher than others. The most popular topics on the hot list include "Events and Performances" and "Discovering New Places."