Email categorization of incoming emails according to the departments using NLP and ML
The client is a renowned and leading bank firm, with its presence all around the globe and more than 80k employees. They provide a wide spectrum of top-notch banking services globally.
Being a leading banking firm, it is no surprise that every single day it receives a gigantic amount of incoming emails for various departments. It became overwhelming for their accountable employees to keep track of the relevant email with business priority.
The client approached us with the requirement to develop a solution that enables filtering and categorizing a mass volume of emails.
Created a data pipeline to clean and pre-process raw chunks of email data. The data is then fed into ML models to determine the output category of the email and forwarded to the concerned employee.
The pipeline contains nodes to standardize email formatting (Removing names, signs, numbers, punctuation) such that the privacy of the emails remains intact without hampering any security compliances.
Leveraged the Keras deep learning API of LSTM to determine the themes (Credit Card, loans, mortgage, etc.) within emails.
Improved email turn-around time by at least 50% for customers
Increased personal productivity of employees. The pipeline allows filtering such that the accountable employee only receives the relevant email.