At JAPAN AI, we’re passionate about using generative AI to tackle real-world challenges. As part of this mission, our team, the JAI GV Team, took part in the 2nd JAPAN AI Hackathon, where we developed an innovative Trip Planner AI Agent. We’re proud to announce that our project achieved the Business Breakthrough Award, recognizing its potential to transform the travel planning experience. In this blog, we’ll dive into the concept, technology, development journey, challenges, and future vision for our award-winning solution.

Concept and Purpose: Simplifying Travel Planning

Travel planning can feel like a daunting task. Between comparing flight prices, checking hotel availability, and balancing preferences like budget and hotel quality within a certain date range, it’s easy to get overwhelmed. Our Trip Planner AI Agent was born out of a desire to simplify this process. The goal? To create an intelligent tool that takes just five inputs: destination, trip dates, flight class, hotel rating, and budget, and delivers personalized flight and hotel recommendations in minutes.

This solution addresses a common struggle: too many options, not enough personalization. By automating the research and decision-making process, we aimed to save users time and reduce stress, making travel planning enjoyable rather than exhausting.

Technology Stack: AI Meets Real-Time Data

To bring our vision to life, we leveraged a powerful combination of AI and APIs:

• Machine Learning & Recommendation Systems: At the heart of our agent is a recommendation engine that filters hotels and flights based on user requirements.
• Natural Language Processing (NLP): NLP helps the agent interpret user inputs, laying the groundwork for future enhancements like understanding more nuanced requests (e.g., “I want a hotel near the beach”).
• Google Travel APIs: We integrated the Google Hotel API and Google Flights API to fetch real-time data on pricing, availability, and schedules. These trusted sources ensure our recommendations are accurate and up-to-date.
• Multi-Tool Workflow: The seamless integration of AI, APIs, and user inputs creates a cohesive system that delivers fast, reliable results.

This tech stack allowed us to build a tool that is both smart and practical, striking a balance between personalization and efficiency.

Development Process: Ingenuity at Work

Our development process followed a clear workflow:

1. User Input Collection: Gather the five key inputs from the user.
2. Data Analysis: Analyze the destination, budget, and time constraints to identify viable options.
3. API Queries: Pull real-time flight and hotel data using Google APIs.
4. Filtering & Ranking: Narrow down options based on user criteria and rank them by price, quality, and relevance.
5. Combination & Recommendation: Pair flights and hotels into cohesive travel packages for the user.