BABA(US:BABA) Huachuang Securities·2025-05-05 08:13Investment Rating - The report maintains a "Recommendation" rating for the online travel industry, expecting the industry index to rise more than 5% over the next 3-6 months compared to the benchmark index [49]. Core Insights - The report highlights the launch of "Wen Yi Wen" by Fliggy, which is the first multi-agent AI application in the domestic online travel sector, designed to provide 24/7 travel planning services [10][11]. - The product utilizes proprietary data and multi-agent collaboration to enhance the travel planning experience, ensuring a seamless end-to-end service from user demand input to transaction completion [10][11]. - The evaluation of "Wen Yi Wen" shows strong performance across five dimensions: accuracy, relevance, data richness, content value, and differentiation, confirming its reliability [10][11]. Summary by Sections 1. Fliggy "Wen Yi Wen": Multi-Agent Driven 24/7 Travel AI Application - Fliggy "Wen Yi Wen" is positioned as a butler-like travel AI application that dynamically responds to user needs, generating executable travel plans and facilitating bookings for flights, hotels, and attractions [10][11]. 2. Core Highlights: Multi-Agent Collaborative Planning - The application features multi-agent collaboration, rich vertical data support, and deep coverage of travel links, which collectively enhance the effectiveness of travel planning services [11]. - The multi-agent team includes roles such as "Itinerary Assistant," "Route Customizer," "Smart Transport Advisor," "Hotel Consultant," and "Strategy Expert," allowing for comprehensive task execution [11]. 3. Functionality Testing: Coverage of Travel Planning, Flight and Hotel Comparison, Destination Exploration (a) Itinerary Planning - The system can package travel plans according to budget constraints, providing detailed itineraries that include route overviews, interactive maps, and essential travel information [22][24]. (b) Flight Comparison - The flight comparison module employs a multi-objective optimization model to generate high-cost performance flight options, significantly improving user decision-making efficiency [32]. (c) Hotel Recommendations - The hotel recommendation feature utilizes multi-dimensional comparisons to match user needs, filtering out hotels with a negative review rate above 10% and providing differentiated options [34]. (d) Destination Exploration - The destination exploration module uses deep semantic analysis to generate personalized recommendations based on user profiles, budget constraints, and real-time data from the supply chain [36].