1. Overview (Project Background)

Hana Tour pursued the adoption of generative AI-based conversational AI Agents to simultaneously strengthen business productivity and service competitiveness in the areas of travel product planning and customer consultation.
The existing travel product planning and consultation process required comprehensive review of multiple internal data, external information, and latest trends, and the structure required significant time and manual work from deriving insights about new destinations to composing package itineraries.

Accordingly, Hana Tour built an AI Agent supporting overall travel itinerary planning and consultation and initiated the project with the goal of assisting travel product planning (MD) work and providing customers with more personalized travel consultation experiences.

2. Challenge (Problem Definition)

Prior to project implementation, Hana Tour faced the following limitations and challenges.

• Limitations of scenario-based chatbots
  The existing chatbot operated primarily on fixed scripts, had low responsiveness to unstructured questions, and required improvements in flexibility and accuracy.

• Inefficiency in utilizing dispersed data
  Internal package product information and external travel trends and local information were scattered, making it difficult to comprehensively utilize them for itinerary planning and consultation.

• High proportion of manual work in travel product planning
  Due to repetitive and time-consuming tasks such as new destination analysis, itinerary composition, and route design, there were limitations in improving MD productivity.

• Lack of personalization in customer consultation experience
  There were constraints in providing hyper-personalized consultation that sufficiently reflected customer reservation history, preferred destinations, and itinerary information.

3. Solution (Resolution Approach)

Hana Tour, together with MegazoneCloud, built an Agentic AI workflow based on Amazon Bedrock and implemented an AI Agent supporting travel product planning and consultation.

• Construction of AI package itinerary design service
  Implemented a work-support AI service that automatically proposes new travel itineraries by comprehensively analyzing package product itineraries and routes actually sold internally, latest travel trends, local information, travel routes, and required time.

• Amazon Bedrock + OpenSearch-based context integration
  Built an agentic AI structure by linking Amazon Bedrock (Claude 3.5 Sonnet) and OpenSearch to comprehensively analyze internal package product data and external trends, and generate itinerary design and consultation responses based on this analysis.

• Implementation of multi-channel and multimodal consultation environment
  Applied STT (speech recognition) and TTS (text-to-speech) to simultaneously support chat and voice-based consultation, providing consultation experiences accessible to diverse age groups and user environments.

• Executable command-based consultation functionality
  Implemented command functionality that recognizes execution intentions such as flight ticket reservations, itinerary changes, and product recommendations during consultation and immediately performs services without requiring separate page navigation.

4. Result (Achievements)

Through AI Agent adoption, Hana Tour achieved the following tangible results.

• Increase in user numbers
  Compared to scenario-based chatbots, users increased by 267% at beta service launch, and expanded to 432% after official launch.

• Improved travel product planning productivity
  By automating the process from new destination analysis to itinerary composition, reduced repetitive work for MDs and created an environment to focus on more strategic planning work.

• Provision of hyper-personalized consultation experience
  By providing customized consultation reflecting customer reservation history and preferences 24/7, simultaneously improved consultation quality, speed, and convenience.

• Expanded user age range and enhanced accessibility
  Through natural language-based interface and voice functionality, built an environment where users with low digital familiarity can easily use the service.

• Established foundation for travel service paradigm shift
  Through AI-based hyper-personalized consultation and itinerary design, secured differentiated cases presenting new service standards in the travel industry.