Ctrip s Technology ABC AI, Big Data, Cloud Computing 02/2018
Ctrip Group at a glance CNY 600B 2017 GMV CNY1.2Trillion 2018 GMV Estimated 520K Domestic hotels 670K International hotels 2 2 150M+ >70% 425 2 International airlines Average Monthly Active 1 Users Booking volume from mobile 3 5,000 Cities connected 2 Note: 1 Average monthly active users based on data for Ctrip, Qunar and Skyscanner combine during the fourth quarter of 2017; 2 Data as of December 31, 2017; 3 Data for the quarter ended December 31, 2017
Ctrip accounts for 65% of Chinese online travel market share Online travel market share in China 0,9% 1,0% 2,4% 13,3% Others 3,1% 6,5% 2 5,1% 65,2% 2 2,5% 1 Source: CTCNN.com and iresearch Note: 1 Market share includes Ctrip and Qunar; 2 Companies in which Ctrip has invested 3
Ctrip s Business scale in China Accommodation Flight Railway Bus & Ferry Package Tours Rental Cars 300M Roomnights 350M Air-tickets 500M tickets 50M Tickets 10M Tourists 1.5M Transactions 60-70% first class and business class are sold by Ctrip 4
A.I Ctrip s Technology ABC AI, Big Data, Cloud Computing 5
Face Recognition First time in the world for an attraction in Fu Jian Province
Smart Check in/out VR room 30s face check in Smart control Check out EASY,direct smarter More comfortable Faster
Phone voice/ IM Inteligent Robot Deep learning:lstm(long Short-Term Memory)+ CNN(Convolutional Neural Network) Understand what customers say Only Long tail questions will be handeled by human operator
Poem creation Deep learning:cnn NLPNeuro-Linguistic Programming Attraction data base Attraction recognition: Red building; Plum blossom; Gender and emotion recogonition: serious; young male;
Poem creation Deep learning:cnn NLPNeuro-Linguistic Programming Attraction data base Attraction Photo: West lake; Timing recognition: Autumn; night; wether;
B.D Ctrip s Technology ABC AI, Big Data, Cloud Computing 11
300M Ctrip Users, generating per day 12
Analysis of traveller demographics Gender, age Means of transportation Spending power 13
Insights on optimal layout of travel resources Avoid duplication of similar travel projects in the same region Promote differentiation of travel resources development Explore new attraction focus Meet diversified customer needs 14
Help build a healthy market Enhance monitoring through technology Point out direction for service improvement Guide customers to choose quality services 15
Precise Forecast Traveller Demographics Internal: 100M/day External: 1.7B/day External Data Internal Data Comprehensive and precise analysis of Customer Demands Travel Trends 16
Personalized Recommendation Customer Behaviors in various business lines Intelligent Computing Personalized Recommendation for Comprehensive Product Offerings Ctrip BU External Partners 17
Customers intention:homepage destination recommendation AB Test: A:Human filter and choose popular destination B:Smart recommendation(based on customers behavior and booking data, different homepage for different page) Smart recommendation: UV click rate 14.9%,2.16 times human working. +115.9% 18
Smart recommendation after one single booking Once flight ticket booked from Shanghai to Madrid: 1 st recommendation will be hotel rooms(strong related) 2 nd : WIFI equipment or SIM card (strong related) 3 rd : Car rental service 4 th : Currency exchange service 5 th : Local tour guide service 19
C.C Ctrip s Technology ABC AI, Big Data, Cloud Computing 20
Cloud Computing Enhance Global User Experience 5 Data centers, 10000+ servers Support globalization Comprehensive services for users from all corners of the world Silicon Valley London Frankfurt Shanghai Singapore Seoul Tokyo Hong Kong Bangkok 21
Cloud call center Customer end Cloud transit Cloud call Operator end PSTN Operator 报表监控 录音质检 Smart IVR VOIP SoftSBC Cloud call center 工单系统 CRM 知识库 Online Chat APP Chat operator Robot Wechat Weibo Cloud service 报表监控 智能质检 移动客服 Email SMS All channels integrated,all data controlled More efficient 工单系统 CRM 知识库
Cloud all channel service: X-Agent 一站式服务体验 全渠道全媒体 X-Agent: PC(Win 应用,Web 应用 ),App Web APP 电话 IM wechat Weibo email 短信
Cloud Call Assign task Batch Call at fixed time customer Business system Auto cloud call supplier Batch Call at fixed time Customized IVR 8788 Application High concurrent call:4000/m, i.e, remind for train tickets payment Voice reminder Stock confirmation prediction
Thanks for your attention!