DvaraAI

Introduction

I have proceeded with Track 3 as I really want to develop agentic ai based systems and also the problem I have chosen is something I have faced recently as I went on to apply for the renewal of my passport. It has been truly exasperating the redundant red tape part of applying for such government-based documents. The project I am working on aims to solve this very issue.

Problem Statement:

Hundreds of Millions of Indians find it hard to obtain government-issued documents like Aadhaar, PAN, and Passports because of the fragmented and complicated systems. Along with having to deal with third-party agents, it leads to exorbitant costs, a lack of clarity in the bureaucracy, and a high chance of corruption. Existing online documents services offer portals for distinct documents, but there is no integration, no customization, no true automation of the entire process. Every user has to untangle multiple websites with shifting requirements, recurrent document submissions, and with very little guidance. All existing solutions are not able to unify the entire user journey and do not provide a single seamless and transparent journey.

Proposed Solution & Unique Features:

DvaraAI is an agentic AI powered digital gateway which aims to unify, automate and personalize the entire document application journey. Unlike bots or portals which work in silos, DvaraAI has standout features, including:

• A single multilingual AI for Aadhaar, PAN and Passport applications.

• AI agents driven autonomous document validation and form filling.

• Proactive error checking and real-time eligibility verification along with deadline alerts.

• Elimination of middlehumans and direct government API integration for reduced delays.

• Action logs capture data for unchangeable records, ensuring transparency and no bribery risk.

DvaraAI stands out for its unique end-to-end methodology that simplifies access, enhances user empowerment, and builds trust.

Tech Stack:
Our solution incorporates the open-source Granite 3.0 LLMs from IBM for understanding, vernacular translation, and form-filling automation; for workflow orchestrations and agent deployment, Watsonx Orchestrate’s Agent Development Kit is utilised; and IBM’s secured APIs are employed for portal integration, scheduling, and notifications.

Scalability & Feasibility:

DvaraAI is designed for modularity and rapid scalability. New services can be added with shallow engineering effort. Additionally, DvaraAI’s cloud-native, containerised infrastructure enables elasticity for large-scale deployment across India’s diverse population.

Technical Architecture:

Agents powered by Granite LLMs endowed with Watsonx orchestration process user interaction. Specialized service agent subroutine calls, document uploads, and feedback using a well secured and auditable backend which unlocks seamless citizen-first services.

By leveraging AI agents for autonomous form-filling and validation, DvaraAI provides a seamless, transparent, and user-empowering experience from start to finish.

Built With

• 3.0
• api
• granite
• watsonx