Signzy Technology : Enabling Real Time and Efficient Bank Account Efficiency
The task of opening a back account when done manually in the legacy system is painfully slow and consumes a lot of business resource as compared to real-time operations. The manual processes also breed other problem areas like faulty data entry due to human negligence, loss of data or fraud because of carelessness or connivance. Eliminating these pain points of the manual system at banks, Signzy is leveraging deep learning based computer vision technology to enable real-time account opening with enhanced compliance adherence, and thorough fraud detection. The company has named its deep learning model as (K)Netra which adds value to the banking operation.
Better Compliance and Fraud Detection
Significantly cutting down on the unnecessary manual processes, the application follows all the relevant steps of collecting required documents, scanning, categorizing, and feeding them into CBS system, but in an efficient way. The documents including CIS and AOF forms are collected and scanned at once saving a lot of time. After scanning, the AI engine does automatic classification of general instruction on the first page of the form using ‘object detection’; it accurately detects face on the customer’s photo using ‘Computer Vision’. For ID detection and signature extraction from the customer’s form, the AI engine uses proprietary algorithm and feature detection, respectively. The application creates the final output in the compatible CKYC form. Along with drastically reducing operation time and other inefficiencies, Signzy’s AI engine helps banks save on manpower costs, improve accuracy and archival.
Seamless Integration and Scalability
The frontend of the application provides an effortless desktop experience to the Teller at bank branch with its easy user interface. It seamlessly integrates with scanners connected to computers across branches and directly scans through them with a single click. Acknowledging the problems of transferring big files in poor connection speed, the application is modeled for optimized compression which allows image compression at capture level, providing full scale utility for branches operating even on the lowest bandwidth.
Better Compliance and Fraud Detection
Significantly cutting down on the unnecessary manual processes, the application follows all the relevant steps of collecting required documents, scanning, categorizing, and feeding them into CBS system, but in an efficient way. The documents including CIS and AOF forms are collected and scanned at once saving a lot of time. After scanning, the AI engine does automatic classification of general instruction on the first page of the form using ‘object detection’; it accurately detects face on the customer’s photo using ‘Computer Vision’. For ID detection and signature extraction from the customer’s form, the AI engine uses proprietary algorithm and feature detection, respectively. The application creates the final output in the compatible CKYC form. Along with drastically reducing operation time and other inefficiencies, Signzy’s AI engine helps banks save on manpower costs, improve accuracy and archival.
Seamless Integration and Scalability
The frontend of the application provides an effortless desktop experience to the Teller at bank branch with its easy user interface. It seamlessly integrates with scanners connected to computers across branches and directly scans through them with a single click. Acknowledging the problems of transferring big files in poor connection speed, the application is modeled for optimized compression which allows image compression at capture level, providing full scale utility for branches operating even on the lowest bandwidth.
Also, its scalable message queues allow handling of massive amount of image files from across all branches in the country. The application raises non-compliance and fraud detection alerts if it finds any illegible, wrong or missing data/documents.
The application works on deep learning algorithms - SSD and CNN architecture. That means it can be trained for more complex tasks by simply getting employees to work on the application thereby training it, enhancing its scalability and efficiency with the ongoing processes. The whole machine learning process is supported through a robust server that also supports connectivity with branches across the country with its Queue based highly scalable infrastructure. Signzy has already registered success leveraging this application for one of the largest PSUs, saving them 645 man resources while adhering to higher compliance with even greater efficiency.
Signzy is leveraging deep learning based computer vision technology to enable real-time account opening with enhanced compliance adherence, and thorough fraud detection. The company has named its deep learning model as (K)Netra which adds value to the banking operation
The application works on deep learning algorithms - SSD and CNN architecture. That means it can be trained for more complex tasks by simply getting employees to work on the application thereby training it, enhancing its scalability and efficiency with the ongoing processes. The whole machine learning process is supported through a robust server that also supports connectivity with branches across the country with its Queue based highly scalable infrastructure. Signzy has already registered success leveraging this application for one of the largest PSUs, saving them 645 man resources while adhering to higher compliance with even greater efficiency.
