Say Hello to AI-KYC: The Next-Gen KYC Verification

ยท

7 min read

Say Hello to AI-KYC: The Next-Gen KYC Verification

Struggling with clunky KYC procedures? AI-KYC is the answer! Our intelligent system simplifies and speeds up verification, turning a complex task into an effortless one. Revolutionize your KYC experience now!

๐Ÿš€ Meet our Team

Hi! Iโ€™m Utsav Bhattarai ๐Ÿ‘‹, and Iโ€™ve teamed up with my friend Rajesh Khadka and our video wizard Bishesh Marasini to bring you AI-KYC. Weโ€™re thrilled to share our latest project with you. AI-KYC is an innovative app designed to make the KYC process smarter and simpler.

Connect with us:

Weโ€™d love to hear from you and show you how AI-KYC is changing the game in identity verification!

๐Ÿ’ก Project Overview

AI-KYC is an innovative web-based application designed to simplify the KYC (Know Your Customer) process for organizations/ Business. It allows orgs to effortlessly create and customize KYC forms directly on the platform. Once KYC forms are set up, the application generates unique links for the respective form that can be shared with end users, who can then fill out the KYC forms with ease. Our machine learning models then work behind the scenes to verify identities by matching document's face with their live selfie.

The platform features a user-friendly dashboard for managing submissions, exporting data to Excel, and viewing key statistics. With JWT authentication, AI-KYC ensures secure and reliable data handling throughout the process.

๐Ÿ”ฅ Why we did this project

In today's business landscape, whether itโ€™s a startup, big company or any organization, Know Your Customer (KYC) verification is essential. However, hiring multiple employees for manual verification (in most of the cases) is costly and inefficient. So, to reduce this problem to some extent, we decided to build this application as a hackathon Project.

Our innovative solution automates the KYC process, allowing businesses and organizations to integrate a powerful AI-driven system that eliminates the need for a large verification team. AI-KYC verifies users in under a minute, reducing costs, enhancing accuracy, and freeing up resources to focus on growth. After investing a little more time on this, we believe we can come up with itโ€™s refined version.

๐Ÿ”ง Tech Tools

๐ŸŽจ Frontend:

  • Vite with React: Fast and optimized development setup using Vite and React.

  • TailwindCSS: Utility-first CSS framework for rapid UI development.

  • Shadcn UI: Comprehensive component library for consistent and customizable UI elements.

๐ŸŒ Backend:

  • MongoDB: NoSQL database for flexible and scalable data storage.

  • Express.js: Web application framework for building robust backend services.

  • Node.js: JavaScript runtime environment for executing server-side code.

๐Ÿค– AI Tools:

  • Gemini Vision API: Extracts details from documents using advanced AI.

  • Face API.js Models: Real-time face comparison between selfies and document photos.

    • 68 Point Face Landmark Detection Model: Identifies key facial landmarks.

    • Face Recognition Model: Matches faces with high accuracy.

    • SSD Mobilenet V1: Object detection model for face localization and identification.

See all the ML modals used in this App here : Click here

โœจNote: Our AI-powered bot works behind the scenes to handle user verification. It uses the above AI models and the Gemini API to analyze form submissions, compare selfies with document photos, and extract essential information for comparison. This ensures a smooth and accurate verification process, keeping everything secure and efficient.

๐Ÿ‘๏ธ Quick Glimpse

1. Hero Page

2. Feature Section

3. Dashboard Page

This page displays the list of forms created by the business organization.

4. Analytics Page

This page provides insights of the forms created by the organization.

5. Setting Page

6. User List Page

This sections shows the list of users who filled the respective KYC form.

7. KYC form Page

This is multi step form where user's fill their details.

๐Ÿ‘ฉโ€๐Ÿ’ผ Architecture Diagram

โœจ Highlighted Features

1. Form Creation:

Businesses and organizations can effortlessly create and customize KYC forms directly on our platform, to make the verification process more easier.

The application automatically generates a unique URL for each KYC form, which businesses and organizations can easily share with their end users. These links can also be integrated into existing systems, making it simple for users to access and complete the forms.

3. AI-Powered Verification

AI-KYC uses advanced machine learning algorithms to automatically verify and reject user submissions, ensuring accurate and efficient processing.

4. Real time Notification on Customers email

Users who fill out the form will receive real-time email notifications with their verification status within minutes. If their KYC submission is rejected for any reason, AI-KYC provides clear and valid reasons for the rejection as a remark.

5. Comprehensive Dashboard

AI-KYC offers a robust dashboard to manage and view KYC form submissions, helping organizations track and analyze responses effectively.

5. Export to Excel

Easily export form data to Excel files for seamless data manipulation and reporting.

6. Statistics Overview

Business/ Organization can gain valuable insights and improve processes with detailed statistics on KYC form submitted by end users.

7. Secure JWT Authentication

Our app uses JWT to protect user data across the server.

8. Modern, Responsive Design

Enjoy a sleek, modern design that is responsive across all devices, ensuring a smooth and user-friendly experience.

๐Ÿ› ๏ธ Challenges

Adapting AI Models

Initially, we aimed to develop and train our own AI models locally. However, this required significant CPU, GPU, and memory resources, which our laptops couldn't handle. We decided to use per-trained models but faced issues with compatibility, particularly on Windows. Migrating to Linux helped, but finding the right models and integrating them was a challenging and frustrating process. Ultimately, we overcame these hurdles and successfully integrated the necessary AI models.

๐ŸŽฅ Video Creation and Presentation

Creating this video was a significant step for us, especially as introverts who faced challenges like multiple retakes and background noise. We wanted to clearly communicate our vision for AI-KYC and how our product can solve real-world problems. Committed to our idea, we believed that simply talking about it wasn't enough; we needed the world to visualize its potential.

A picture is worth a thousand words, so we chose to express our concept through a video. With the help of our friend Bishesh Marasini, who handled the video editing, we were able to bring our ideas to life and show the real-world application of AI-KYC. This video is not just an explanation but a visualization of our dedication and the impact we aim to make.

This pic reflects the video editing process.

๐Ÿ”ฎ Future Plans

1. Seamless Integration:

We plan to integrate AI-KYC directly into existing systems, making it easier for businesses to adopt and use our solution without disrupting their current workflows.

2. Enhanced Machine Learning Accuracy:

We aim to fine-tune our machine learning models to improve accuracy and similarity detection, ensuring even more reliable KYC verification.

3. Providing Verified Certificates or badge:

In future, we are planning to offer verified certificates to customers who complete the KYC process, enabling them to use these certificates across various platforms, streamlining their verification experience.

4. API Outsourcing:

Our goal is to develop and provide an API so that businesses can use AI-KYC in their own unique ways, leveraging our technology to enhance their existing processes.

๐Ÿš€ Deployment

Yes, you heard it rightโ€”AI-KYC is now live! We're using vercel to deploy the front end and render.com to deploy the server and bot components of the application.

bot running in the server

๐Ÿ”— Key Links

๐ŸŽ‰ Wrap-up

Overall, developing AI-KYC was an incredible learning experience. We gained valuable insights into advanced AI models, honed our skills in integrating machine learning algorithms, and improved our proficiency with the MERN stack. And most important of all, we developed a new skill of being patient when our code doesn't work and we don't get any solutions at all. The entire journey was rewarding, and we highly encourage anyone interested in innovation to participate in a hackathon and push their limits.

ย