Know Your Customer, better known as KYC, is a mandatory customer onboarding process for banks and financial institutions to know about their clients’ identities, their information and verify them accordingly. And what the eKYC does is bringing the whole pen & paper formalities to the digital platform; allowing anyone with computer device or a smart phone to sign up seamlessly.
Before eKYC, the traditional bank account opening formalities were paper-based and often took more than one day. And being paper-based, it was very difficult to verify and validate new customers. Hence, no single individual could open a new account at the same time they have submitted all required information.
Challenges
In the traditional sense, the customer onboarding process is done manually. So, customers were required to go to banks & financial institutions physically and do paper-based on onboarding formalities. The number of verification documents is usually quite big and the onboarding process itself is time consuming. Also, it was a hassle for banks as they must manually do the risk assessment and verification process. And the COVID-19 pandemic has forced all sorts of institutions to adopt digital service capabilities, forcing banks and financial institutions to employ online onboarding and assessment solutions.
Solution
With eKYC, now banks & financial institutions can onboard clients easily, and from the comfort of their own homes. The users can submit the digital versions of their documents on their devices and the eKYC will start the onboarding process from there. For collecting NID data we have developed an optical character recognition feature, or OCR. Our OCR can validate real NID images like smart-card NID, old NID, as well as downloaded NID from Election Commission. In this step customer provide us their NID’s front image and back image, where the OCR will read the data from those images in both English and Bangla. In the backend all NID data will be preserved in a specific format. For biometric verification, face verification has been completed in multiple step verification. Initially the liveliness of customer image is verified. This is done through the means of Amazon Rekognition, a machine learning solution from AWS. Liveliness verification is done by customer eye blinking as well as paling customer eye in a predefined rectangular box. If liveliness check verifies, then customer image is verified with NID image data. For further assistance, fingerprint verification can provide a faster and more accurate way to validate customer identity. This directly verifies the customer fingerprint with Election Commission customer fingerprint. And for security measures, there’s data encryption, forgery detection, face spoofing and implementation of standard OAuth2 protocol to secure all calls between mobile application and internal bank services. There is systematic risk grading and the stored information on the eKYC can later be used to prevent any money laundering activities. And banks and financial institutions have a secured database of their clients.
Impact
The eKYC has allowed seamless access to banking services for users. Starting from onboarding, users are not required to go to banks or financial institutions to sign up for their services. Moreover, now users can opt for other banking services like opening new fixed deposits, savings accounts and loans without much issue as all the personal information is stored in the eKYC and can be used to open further accounts and services. And for the banks and other financial institutions, there’s no manual paperwork and time to onboard new customers have been drastically decreased. Their customers are more engaged as they do not have to submit extra documents or verification materials. More importantly, these help banks and financial institutions to cater their services on the digital platform, making accessing the services and accounts for their clients easier.
Features
- NID Text Data Extraction
- Aged face verification with EC (Election Commission)
- Guided Data Collection
- Legal Document Tracking
- Multi-lingual Support
Key Technologies
- Java
- Python
- AWS Face Rekognition
- Google OCR
- Tesseract
- TensorFlow
Project Highlights
