WorldCup2026Conceirge

Inspiration

World Cup 2026 will bring fans from all over the world to cities they may not know well. Planning a match-day trip can be confusing because fans need match details, hotels, fan zones, food, transport, safety tips, and local guidance from many different sources. We wanted to build something that feels like a personal concierge for every fan.

What It Does

World Cup 2026 AI Fan Concierge helps fans plan their trip through one simple chat. A user can ask about a city, match, budget, travel mood, hotel needs, fan zones, or local tips. The app uses multiple AI agents to return a complete plan with match information, hotel recommendations, fan zones, restaurants, transport guidance, and city advice.

How We Built It

We built the frontend with Next.js and created a polished chat interface with streaming responses, mood selection, language control, city selection, and result cards. The backend uses Node.js and Express. MongoDB Atlas stores match data, hotels, fan zones, restaurants, city guides, sessions, and agent results.

We also used MongoDB Vector Search so the app can search by meaning instead of only exact keywords. For example, a user can ask for a “party hotel with World Cup energy,” and the system can find hotels with a similar vibe.

MongoDB And MCP

MongoDB is the core of the project. We first set up MongoDB Atlas, created the database, seeded collections, and created the structure for vector search. We also configured MongoDB MCP and tested it with Codex to confirm that the AI development environment could connect to MongoDB, inspect collections, and verify the database setup.

Challenges

One challenge was connecting all parts together: frontend, backend, MongoDB Atlas, vector search, AI APIs, and deployment. We also faced API rate limits while generating embeddings and responses, so we added fallback logic to keep the app usable even when external services were limited.

Another challenge was making the app feel like a real product, not just a chatbot. We improved the UI with city maps, animated responses, mood selection, language options, and structured cards.

What We Learned

We learned how powerful MongoDB can be for agentic applications. It is not only a database for storing records; it can support semantic search, session memory, structured agent results, and real-time product experiences. We also learned how multi-agent systems can make AI apps more useful by separating responsibilities between specialized agents.

What's Next

Next, we would add real booking integrations, live FIFA data, map APIs, user accounts, payment options, and stronger multilingual support. This could become a real AI concierge platform for sports events, concerts, conferences, and large tourism experiences.