Agentic City Problem Solver

AI-powered city issue management system

Comment

Inspiration:

Cities face countless daily issues such as garbage overflow, road damage, water leakage, and emergency situations. We noticed that most complaint systems are slow, manual, and lack transparency-citizens don’t know who is handling their issue or when it will be resolved. Inspired by the idea of agentic AI systems and smart cities, we wanted to build a solution where AI acts like a smart city officer-understanding problems, assigning responsibility, and tracking progress automatically.

What it does:

  • Allows citizens to report city issues in natural language
  • Automatically analyzes complaints using AI
  • Assigns the issue to the correct department
  • Determines urgency and resolution time (ETA)
  • Generates suggested solutions and action steps
  • Tracks complaints in real time via an admin dashboard
  • Both citizens and administrators see consistent, real-time data including urgency, ETA, and status.

How we built it

Frontend: React.js with a modern, clean UI Backend: Node.js + Express.js AI Intelligence: Google Gemini API

  • Complaint understanding
  • Department routing
  • Solution generation Database: MongoDB Atlas
  • Stores complaints, ETA, SLA, and status Logic Layer:
  • Urgency-based ETA calculation
  • Centralized “single source of truth” for timing

We designed the system so the same data is shared across the complaint form and admin dashboard, ensuring consistency.

Challenges we ran into

  • Handling inconsistent ETA values between different UI panels
  • Designing a clean data flow between AI output and database storage
  • Managing MongoDB Atlas SSL and permission issues
  • Ensuring Gemini outputs didn’t break the system when input was missing
  • Keeping the UI modern yet easy to understand
  • Each issue required careful debugging and architectural refinement

Accomplishments that we're proud of

  • Successfully integrated Google Gemini for real-time complaint intelligence
  • Built a fully working end-to-end system during the hackathon
  • Achieved accurate department routing and ETA calculation
  • Created a clean, professional admin dashboard with SLA tracking
  • Ensured transparency between citizen and admin views

What we learned

  • How to design agent-based AI workflows
  • Best practices for AI-safe backend handling
  • Managing real-time data consistency across multiple UI components
  • MongoDB Atlas deployment and security handling
  • Importance of UI/UX clarity in civic tech applications

What's next for Agentic City Problem Solver

  • Image-based complaint analysis (photos of issues)
  • Notification system (SMS / email updates)
  • Priority escalation for unresolved complaints
  • Department-wise analytics and performance tracking
  • Mobile-friendly version for wider adoption

Built With

Share this project:

Updates