🌟 Inspiration
India’s cities handle thousands of civic complaints daily—potholes, garbage, water leaks, power outages—but resolution is often delayed due to fragmented systems, manual triaging, and lack of transparency. We wanted to build an AI-powered solution that bridges the gap between citizens and authorities, making urban governance faster, smarter, and more accountable.
💡 What it does
NagarSathi is an agentic AI civic issue resolver that automates the entire complaint lifecycle:
Ingests complaints from WhatsApp, Twitter, and municipal portals.
Uses IBM Granite models to understand multilingual complaints and classify them (roads, sanitation, water, electricity).
Employs IBM Agent Development Kit (ADK) to orchestrate workflows: intake → categorization → ticket creation → routing → tracking → citizen updates.
Provides dashboards with real-time complaint heatmaps and trend analytics for municipal decision-makers.
Result: faster complaint redressal, transparent tracking, and improved citizen trust.
🛠 How we built it
Data Ingestion: Connected APIs for WhatsApp, Twitter, and sample municipal complaint datasets.
Granite Models: Used for natural language understanding, summarization, and classification of complaints in English + Indian vernacular languages.
ADK Workflow: Designed end-to-end pipeline for ticket generation, routing, and continuous citizen updates.
Visualization: Built dashboards showing complaint categories, resolution times, and geospatial hotspots.
🚧 Challenges we ran into
Handling multilingual input from diverse citizens.
Designing seamless integration with municipal systems that vary across cities.
Ensuring scalability and explainability of AI decisions.
🏆 Accomplishments that we're proud of
Proved that end-to-end civic automation is possible using Granite + ADK.
Created a multilingual complaint resolution system relevant to India’s diverse population.
Designed dashboards that enable data-driven policy decisions for city authorities.
📚 What we learned
How agentic AI systems can go beyond chatbots and deliver action-oriented workflows.
Importance of Granite + ADK combo for building scalable enterprise-grade AI agents.
Value of India-first design thinking—building for vernacular languages, fragmented systems, and local governance needs.
🚀 What's next for NagarSathi – AI-Powered Civic Issue Resolver Agent
Voice integration so citizens can call and register complaints via speech in local languages.
Predictive analytics to detect hotspots and suggest preventive maintenance before complaints arise.
Partnerships with municipalities to pilot NagarSathi in one Indian smart city.
Long-term: Build a state and national-level governance dashboard powered by aggregated complaint data.
Log in or sign up for Devpost to join the conversation.