Inspiration

University students receive dozens of fragmented messages every day: emails from professors, LMS notifications, deadlines, event reminders, and group chat updates. Important tasks are often buried inside long, confusing messages.

We noticed that most AI tools act like chatbots, but they don’t actually help students make decisions. We wanted to build an AI system that does more than generate text — one that helps students understand what matters and what to do next.

That idea became YoCampus: an AI decision assistant for campus life.


What it does

YoCampus takes a student’s natural language message and converts it into a structured decision result.

Instead of just replying with text, the system:

  • Detects the user’s intent
  • Summarizes the situation
  • Estimates confidence
  • Suggests clear next actions

For example, a message like:

“Professor emailed me about homework due tonight.”

Will be turned into a structured output such as:

  • Intent: urgent academic task
  • Summary: Homework due tonight
  • Suggested action: Complete assignment immediately

The goal is to transform messy information into clear, actionable decisions.


How we built it

The system is designed as a simple AI decision pipeline:

  1. Frontend (React + Vite)

    • Text input interface for user messages
    • Displays AI-generated summary, intent, and suggested actions
  2. Backend AI Decision Engine

    • Receives raw user text
    • Composes a structured prompt
    • Sends it to the LLM (Gemini API)
    • Parses the response into a decision object
  3. Decision Output Structure

    • Summary
    • Intent
    • Confidence score
    • Suggested actions

The frontend communicates with the backend through a REST API, creating a clean separation between UI and AI logic.


Challenges we ran into

One of the biggest challenges was defining the role of AI in the system.

Instead of letting the model generate free-form responses, we wanted it to produce structured, machine-readable decisions. This required:

  • Careful prompt design
  • Enforcing a consistent output format
  • Handling unpredictable LLM responses

Another challenge was time constraints. We had to focus on the core decision pipeline rather than building a full campus app.


Accomplishments that we're proud of

  • Built a complete end-to-end AI decision pipeline
  • Designed a structured decision output instead of a simple chatbot
  • Created a clean, modular frontend architecture
  • Demonstrated a practical AI use case for real student problems

Most importantly, we proved that AI can be used not just to talk, but to help people make decisions.


What we learned

During this hackathon, we learned:

  • How to design AI systems around structured outputs
  • The importance of prompt engineering for reliability
  • How to connect a modern React frontend to an AI backend
  • How to scope a project realistically within hackathon time limits

We also learned that good AI products are less about fancy models and more about solving real user problems.


What's next for YoCampus AI Decision Assistant

In the future, we plan to expand YoCampus into a full campus productivity assistant:

  • Email and LMS integration
  • Automatic deadline tracking
  • Smart reminders and scheduling
  • Personalized decision support
  • Multi-step action automation

The long-term vision is an AI system that not only understands campus information, but also helps students act on it with confidence.

Built With

Share this project:

Updates