FIFA 2026 Match Day AI Planner

Inspiration

• The Global Scale of 2026: With the 2026 World Cup spanning three vast countries (USA, Canada, Mexico), the sheer logistics of planning travel, accommodation, and match schedules are overwhelming.
• A Fragmented Fan Experience: Fans usually have to juggle multiple apps for tickets, maps, hotels, and team updates.
• The Power of AI: We wanted to harness the latest in Generative AI to create a centralized, conversational "super-concierge" that handles the heavy lifting of trip planning.

What it does

• Intelligent Trip Planning: Generates personalized itineraries based on user preferences, budget, and desired matches.
• Context-Aware Q&A: Answers complex questions about stadiums, host cities, and schedules using up-to-date data.
• Seamless Logistics Assistant: Helps users map out the logistics of moving between games in different cities or countries.

How we built it

• Backend & Data: Used Elasticsearch as a robust vector database to store and quickly retrieve massive amounts of World Cup and city data.
• AI Integration: Leveraged the Model Context Protocol (MCP) combined with Google Gemini (@google/genai) to give the AI agent secure, structured access to our custom data and logic.
• Frontend: Built a responsive, dynamic user interface using React and Vite to provide a smooth chat and planning experience.

Challenges we ran into

• Data Aggregation: Finding and structuring reliable data for all the spread-out host cities and stadiums.
• AI Hallucinations: Ensuring the AI planner provides factually accurate schedules and distances rather than making up logistics.
• MCP Integration: Bridging the gap between the generative model and our Elasticsearch backend securely using the relatively new Model Context Protocol.

Accomplishments that we're proud of

• Working AI Agent: Successfully building an autonomous agent that can reason through multi-city travel constraints.
• Fast Retrieval: Implementing highly efficient semantic search with Elasticsearch so the AI gets exactly the context it needs in milliseconds.
• Clean Architecture: Establishing a scalable separation of concerns between the React frontend, backend APIs, and the MCP server.

What we learned

• Agentic Workflows: Deepened our understanding of how to use MCP to expand an LLM's capabilities far beyond its training data.
• Prompt Engineering for Logistics: Learned how to structure system prompts to force the AI to respect strict constraints like travel time and match overlaps.

What's next for FIFA 2026 Match Day AI Planner

• Enhanced Data Correctness: Implement rigorous automated checks and real-time API integrations to guarantee $100\%$ accuracy for match times, flight delays, and sudden venue changes.
• In-App Navigation: Build turn-by-turn routing directly into the app to guide fans from airports to hotels and straight to their stadium seats.
• Direct Calling Features: Add one-tap calling functionality to immediately connect users with local hotels, hospitals, and emergency services in the host cities.
• Live Booking & Transactions: Evolve from a planner to an all-in-one booking platform, allowing users to secure flights, reserve hotels, and buy match tickets directly through the AI chat interface.

Built With

• elastic
• gemini
• mcp
• node.js
• react
• vite