Bridging the Digital Divide: How We Built India's First Government Scheme WhatsApp Bot

When the Tirunelveli district administration launched the GovTechThon challenge, we saw an opportunity to solve a fundamental problem in public service delivery: the accessibility gap between government schemes and eligible citizens. In this article, I'll walk through our journey of creating MakkalNalamBot, India's first WhatsApp-based virtual assistant for government schemes, and share the technical challenges we overcame along the way.

The Problem: Information Accessibility in Public Services

Government schemes are designed to support citizens, but navigating the bureaucracy to find relevant programs remains a significant hurdle. Many individuals, particularly in rural areas, spend countless hours standing in queues at government offices just to learn which programs they qualify for. This inefficiency affects both citizens and administrators.

During my internship at Tirunelveli Collectorate, I witnessed this challenge firsthand. Working across multiple government departments, I realized that a technological solution could bridge this gap effectively.

Our Solution: MakkalNalamBot

Our team of three—RS Siva Subramanian, T Venkadesh, and myself (R Ajithkumar)—envisioned a conversational interface that could guide citizens to relevant government schemes based on their demographic profile, without compromising privacy or requiring technical expertise.

Key requirements included:

• Making information accessible in Tamil language
• Creating a user-friendly, touch-based interface
• Ensuring no personal data storage
• Providing scheme eligibility information in 6-7 interactions
• Directing users to the appropriate department contact points

Technical Architecture and Implementation

#### The Technology Stack

We built MakkalNalamBot using:

• Node.js: For the backend server and business logic
• JavaScript: For core application development
• JSON: For structured data storage and scheme information mapping
• WhatsApp Business API: For message delivery and interaction
• MongoDB: For efficient data organization and retrieval (planned for future expansion)

#### Development Journey

Our project evolved through several stages:

1. Proof of Concept on Telegram: Siva Subramanian, our UK-based chatbot specialist, created an initial prototype using Telegram's API.

1. Data Collection and Structuring: During my internship at Tirunelveli Collectorate, I collected, categorized, and formatted information on 132+ government schemes from eight departments, including the Adi Dravidar and Tribal Welfare department and BC MBC Welfare department.

1. Migration to WhatsApp: We transitioned from Telegram to WhatsApp Business API to reach a wider audience, as WhatsApp has significantly higher penetration in rural India.

1. Building the Conversation Flow: We designed a decision-tree architecture where users navigate through options by selecting their gender, age, and income range through simple touch interactions rather than typing.

#### Technical Challenges and Solutions

Challenge 1: Structuring Diverse Scheme Data
We needed to organize varied scheme information in a standardized format that could be easily queried.

Solution: We created a hierarchical JSON structure with schemes categorized by department, eligibility criteria, and benefits. This made the data easily searchable and extensible:

javascript
{
  "schemes": [
    {
      "id": "scheme-001",
      "name": "Pre-matric Scholarship",
      "department": "Adi Dravidar and Tribal Welfare",
      "eligibility": {
        "gender": ["male", "female"],
        "age": { "min": 6, "max": 18 },
        "income": { "max": 200000 }
      },
      "benefits": "Financial assistance for education",
      "contactPoint": {
        "office": "District Welfare Officer",
        "phone": "04622-XXXXXX"
      }
    }
  ]
}

Challenge 2: Privacy Concerns
We needed to provide personalized recommendations without storing personal data.

Solution: We implemented a stateless architecture where user inputs are processed in-session without persistence. No Aadhaar or personal identifiers are stored, and previous interactions are automatically purged after completion.

Challenge 3: WhatsApp Business API Integration
Working with WhatsApp Business API presented unique challenges compared to Telegram.

Solution: We developed middleware that handled message formatting, interactive buttons, and session management specific to WhatsApp's requirements. This included:

javascript
// Sample code for handling WhatsApp interactive messages
function createSchemeSelectionMessage(schemes) {
  const buttons = schemes.slice(0, 3).map(scheme => ({
    type: "reply",
    reply: {
      id: scheme.id,
      title: scheme.name.substring(0, 20)
    }
  }));
  
  return {
    type: "interactive",
    interactive: {
      type: "button",
      body: {
        text: "You may be eligible for these schemes:"
      },
      action: {
        buttons
      }
    }
  };
}

Challenge 4: Language Localization
Supporting Tamil was essential for reaching the target audience.

Solution: We implemented a translation layer that mapped all user interface elements and scheme information to Tamil. This was crucial for ensuring accessibility for non-English speakers.

Impact and Results

Our MakkalNalamBot won first place among 500+ teams at GovTechThon, earning us a ₹1 lakh prize and official recognition from the Tamil Nadu IT Minister. More importantly, the project was adopted by the district administration and is being scaled to include more schemes.

Key metrics:

• 132+ government schemes integrated
• Support for 8+ government departments
• 6-7 touches to get relevant scheme information
• Zero personal data storage

Future Development Plans

We're actively working on several enhancements:

1. Expanding to Central Government Schemes: Adding national-level programs to the database
2. Multi-language Support: Including more regional languages beyond Tamil
3. Advanced Analytics: Implementing usage patterns analysis to improve user experience
4. Integration with Government Portals: Creating direct application capabilities
5. Personalized Notifications: Alerting users about new schemes they may qualify for

Conclusion

MakkalNalamBot demonstrates how thoughtful application of technology can transform public service delivery. By leveraging familiar platforms like WhatsApp, we've created a solution that meets citizens where they are, eliminating barriers to accessing vital government programs.

The project exemplifies how modern development practices and accessible technologies can drive meaningful social impact. As we continue developing this platform, our vision is to create a model that can be replicated across states and eventually nationwide, ensuring every citizen can easily access the government support they need and deserve.

About the Author: R Ajithkumar is a software developer who interned at Tirunelveli Collectorate and was instrumental in data collection and technical development of MakkalNalamBot. He specializes in conversational interfaces and public service automation.