Ø Intend of the project to develop a paradigm using conceptualization of Signal Processing

and Advance Machine Learning Algorithms to identify Fraudulent and Genuine Calls from

alerted audio calls of HSBC Contact Centre.

Ø Implemented signal processing and Channel Diarization concepts to segregate the

customer conversations from mixed raw signal. Developed scripts to derive features from

audio signal by reminiscing behaviors of fraudsters and genuine customers from audio and

implemented into Machine Learning models to predict Voice labels.

Ø Plan to replicate Fraud Ops Manual caller verification process by integrating the developed

paradigm which would impact the reduction of Fraud losses approx. 1.1M GBP monthly

Ø An automated business assistive system enables customers to know their application status

(Approved/Cancelled/Declined), amount of time left to complete the verification process

of application, amount of time it will take to final approval of application.

Ø Assists Operation AMOs/MOs, stake holders, etc. to know monthly Application volumes,

KPIs, Application Status, Application Cancellation Reason, etc.

Ø Understanding the verification process of Cards, Loans, Mortgages, and Other products for

multiple markets. Identifying TTD and Referred volumes based on Captured date,

Decisioned date and other parameters.

Ø Forecasting TTD AOP volume and Referred AOP volume using multiple forecasting

methods.

Ø Calculating Staff Productivity, RPH, etc.

Ø Predicting Staff Demand using Machine Learning models for multiple markets