On July 17th, Ctrip held a press conference in Shanghai and officially launched its vertical large-scale model product “Ctrip Wendao”. According to James Liang, Chairman of the Board of Directors of Trip.com Group, Ctrip Wendao is the first vertical large-scale model developed by Ctrip itself in the tourism industry, with model parameters exceeding billions.
In the context of intelligent society, the tourism industry, as a “spiritual demand industry that is difficult to automate,” will see an increase in demand with the overall improvement of social prosperity, and its contribution to the economy will also expand. “Ctrip will actively embrace the intelligent society, focusing on the global tourism industry and utilizing AI to provide intelligent assistant services before, during, and after travel,” said James Liang.
According to the information, the “Ctrip Wendao” released this time currently has two main capabilities:
Firstly, it can provide services for querying and guiding bookings based on users’ explicit needs. Users can directly send text or long natural language voice messages to make flight, hotel, and other product reservations.
Secondly, it offers intelligent travel recommendation services for users with undecided needs. When users express their ideas, Ctrip Wendao can recommend travel destinations, hotels, attractions, itinerary planning options, and discounted booking choices.
In order to solve the problem of general AIGC, Ctrip has invested heavily in building a “reliable answer database” for the tourism industry. Based on intelligent algorithms, Ctrip manually verifies and forms the “Ctrip Reputation Ranking” for commonly recommended themes related to hotels, attractions, and itineraries. However, due to the close correlation between travel and hot trends, users also hope to avoid price peaks. As a result, the “Ctrip Hot Ranking” and “Ctrip Specials Ranking” were born. This group of rankings has become representative content products with distinctive features from Ctrip.
James Liang said, “The average data calculation volume per ranking list reaches 5 million, and after being verified by employees from over 30 countries and regions worldwide, it is the cream of the crop for users.”
Currently, Ctrip’s large-scale model has started online internal testing. After its official launch, it will meet various travel-related questions from users, such as destination consultation and flight and hotel booking, etc.
In fact, the reason why Ctrip was able to be the first in the travel industry to release a vertical large-scale model is closely related to its continuous investment in digital technology.
According to the information, from 2018 to 2022, Ctrip’s cumulative investment in product research and development expenses amounted to about $6.4 billion. In the challenging year of 2020, which was heavily impacted by the epidemic, the proportion of Ctrip’s operating expenses dedicated to product research and development increased by 10% compared to before the outbreak. During that same year, within Ctrip’s employee system, the product research and development team had the largest scale with a team member ratio reaching 48%. By 2022, its product research and development expenses accounted for 42% of its net income.
Under such a significant investment in research and development, Ctrip has long transformed from a travel platform to a technology company, driven by technology for business growth.
In addition, Ctrip has established call centers covering multiple countries to provide continuous service to users through tens of thousands of customer service representatives and a 24/7 global unified service standard. Multilingual AI customer service will be an important supplement to Ctrip’s global service capabilities. Multilingual services have become a major driving force for the growth of Ctrip’s overseas business. Taking the hotel department as an example, from 2023 until now, the volume of inquiries received by Ctrip’s overseas customer service has nearly quadrupled compared to the previous year.
In actual application, in the first quarter of 2023, Ctrip’s overall hotel bookings on the international platform reached a historic high, more than double the booking level in 2019. AI technology has been deeply integrated into three major scenarios of Ctrip customer service, saving over 10,000 hours of customer service work time per day and freeing up more than 1,000 Ctrip customer service personnel daily. The self-service problem resolution rate for users has doubled. In the online consultation scenario, the overall self-service rate of text-based chatbots has increased to around 75%.
For international business scenarios, the language coverage of chatbot customer service has significantly expanded from six languages to twenty-four languages. The self-service rate has also increased from less than 60% to approximately 75%, similar to that of Chinese-language services.