AdminAI

• 

  AI-powered civic platform for managing citizen complaints, prioritization, and government workflows.

• 

  Citizen Complaint Submission – Simple form for submitting civic complaints with details and supporting images.

• 

  AI-powered governance platform for citizen complaint management, classification, priority detection, and analytics.

• 

  AI Complaint Analysis – Automatically identifies the department, category, confidence score, summary, and priority.

• 

  Admin Dashboard – Centralized complaint management with search, filters, status tracking, and complaint overview.

• 

  **Complaint Status Distribution – Visual overview of pending, in-progress, and resolved complaints.**

• 

  Complaint Details – Displays citizen information, AI classification, confidence, priority, status, AI summary, and officer remarks.

• 

  AI Analytics Dashboard – Provides complaint intelligence through status tracking and department-wise analytics.

• 

  **AI Category Distribution – Visualizes how complaints are classified across different categories using AI.**

• 

  **AI Insights – Highlights complaint volume by department and provides AI-driven recommendations for faster resolution.**

• 

  **Officer Login – Secure department-based access for government officers to manage assigned complaints.**

• 

  **Officer Dashboard – Department-specific portal where officers view and manage assigned citizen complaints.**

• 

  **Complaint Update – Officers can track status, add action remarks, and update complaint progress.**

• 

  **AI Summary & Officer Action – AI-generated complaint summary with officer remarks(after updated by officer) and submission tracking.

• 

  Officer updates are reflected in real-time analytics.

Inspiration

Citizen complaints about roads, sanitation, public services, and other civic issues can be difficult to manage when they arrive in large numbers. Manual sorting and prioritization can take time and may delay action.

We wanted to build an AI-powered platform that could help administrators and government officers understand complaints faster, identify their departments, prioritize urgent issues, and manage the resolution process more efficiently.

What it does

AdminAI is an AI-powered civic intelligence platform that provides an end-to-end complaint management workflow.

• Citizens can submit complaints through a simple interface.
• AI analyzes complaint descriptions and predicts the appropriate category.
• The system provides AI confidence and complaint priority.
• AI generates concise summaries to help officers understand complaints quickly.
• Administrators can monitor complaints through a centralized dashboard.
• AI Analytics provides insights into complaint trends.
• Department officers can securely log in and view complaints related to their department.
• Officers can update complaint status and add remarks about actions taken.

How we built it

We built AdminAI using Python and Streamlit for the application interface, with SQLite for database management.

For AI-based complaint classification, we used Natural Language Processing with TF-IDF feature extraction and a machine-learning classifier. The system integrates the trained model into the complaint workflow so that submitted complaints can be analyzed automatically.

The application is organized into separate components for the user interface, database operations, machine-learning prediction, analytics, and officer management.

We used Git and GitHub for version control and deployed the application using Streamlit.

Challenges we ran into

One of our main challenges was integrating the machine-learning model with the database and web application while keeping the complete complaint workflow reliable.

Another challenge was making the AI output useful for actual users. Instead of showing only a predicted category, we added confidence, priority, and AI-generated summaries to provide officers with more useful information.

We also had to design separate workflows for administrators and department officers while maintaining a simple user experience.

Accomplishments that we're proud of

We are proud of building a complete working prototype rather than only an isolated machine-learning model.

AdminAI connects citizen complaint submission, AI classification, priority prediction, AI-generated summaries, administrator analytics, and department-level officer workflows in one platform.

We are especially proud of integrating the AI predictions directly into the officer workflow, allowing officers to understand complaints quickly and update their status and remarks from the same portal.

What we learned

This project helped us understand how machine learning can be integrated into a practical real-world application.

We learned how to work with NLP and TF-IDF, integrate trained machine-learning models into Streamlit, manage application data using SQLite, design role-based workflows, and connect AI predictions with a useful user interface.

We also learned that building an AI application involves more than model accuracy—it requires designing the complete workflow around the people who will use the system.

What's next for AdminAI

We plan to expand AdminAI with:

• Tamil and English multilingual complaint understanding
• Voice-based complaint submission
• Improved urgency and priority prediction
• Geographic complaint hotspot analysis
• Automated complaint escalation
• SLA and resolution-time tracking
• Citizen complaint tracking
• Advanced analytics and reporting
• Stronger security and audit trails

Our long-term goal is to evolve AdminAI into a practical civic intelligence platform that helps public-service departments respond to citizen complaints faster and more effectively.

Built With

• machine-learning
• natural-language-processing
• python
• scikit-learn
• sqlite
• streamlit