Shayak

1 line = problem+solution+Ai role

Comment

Inspiration

In Indian cities, citizens often notice civic problems such as broken roads, overflowing garbage, water leakage, or damaged streetlights, but do not know where to report them. Even when complaints are submitted, they may be poorly described, difficult to prioritize, or routed to the wrong department.

We wanted to build an AI agent that makes civic reporting simple for citizens and actionable for authorities.

What it does

Sahayak is a multilingual civic complaint agent. A citizen can report an issue using text, a photo, or a voice note.

The app uses Gemini to understand the complaint, identify the issue category, estimate urgency, extract location details, and generate a structured civic ticket. The complaint is then displayed on an officer dashboard with a priority queue, map view, and suggested next actions.

Key features:

  • Report issues through text, photo, or voice
  • Gemini-powered complaint understanding and classification
  • Automatic urgency and department routing
  • Google Maps-based location view
  • Duplicate complaint detection
  • Officer dashboard for prioritized issue management
  • Multilingual support for Hindi, Marathi, and English

How we built it

We built the frontend using Next.js/Flutter, authentication and data storage using Firebase, and AI reasoning using the Gemini API. Google Maps Platform is used for location visualization and geospatial context.

When a user submits a complaint, Gemini analyzes the input and returns structured data such as issue type, severity, location, summary, and recommended action. This data is stored in Firestore and shown to municipal officers through a dashboard.

Challenges we ran into

One challenge was converting unstructured citizen inputs—such as blurry photos or informal voice notes—into reliable, structured complaint data. We solved this by designing clear Gemini prompts, adding validation steps, and allowing users to confirm or edit the extracted details.

Another challenge was making the interface simple enough for users with different levels of digital literacy.

Accomplishments that we're proud of

We built a working end-to-end prototype where a citizen can submit a real-world issue and an officer can see a prioritized, location-aware ticket. We are especially proud that the project uses AI not just for conversation, but to complete an actionable workflow.

What we learned

We learned how to design an AI agent around a real operational workflow, how to combine Gemini with Firebase and Google Maps, and how important prompt design and user confirmation are for reliable AI outputs.

What's next

Next, we plan to add WhatsApp integration, automatic escalation for high-priority issues, public complaint-status tracking, and analytics for municipal departments.

Built With

Share this project:

Updates

Submission history