Automated Email Classification for one of the World’s Largest Retailers

Industry: Retail


Our client’s team goes through a large number of emails manually on a daily basis. They need to categorize them as per business logic and respond within 24 to 48 hours to maintain their service level agreement. Manual post-analysis is done on monthly basis to generate reports. Following were some challenges in the process –

  • During the daily support system, IT Team receives huge number of emails from customers and also receives feedback in structured and unstructured format
  • It has been a semi-automated process in terms of data management and communication to their suppliers
  • Those emails are critical to address and the typical response time is 24 – 48 hours considering the criticality of the issues.
  • To meet the SLA, the team required was increasing to handle the volume and the process is not scalable and cost effective.
  • So, an automated email categorizer was needed to help the support team to bring efficiency in the overall process

Project Journey

The heavy rise in the volume of received emails was translating into an increase in the number of personnel required to handle the issue; proving to be cost-ineffective. Our client was in need of an automated email categorizer to help their communication experts and bring efficiency on board.


While Xpresso has been used as core engine to categorize email automatically through natural language understanding, lot of other application customization has been done to make it fully functional.

  • Each & Every Email for this purpose, is polled from Email server to Email Categorization Database, through customized Email Fetcher and Email Parser
  • All these emails will be categorized in below 2 levels –
    • Application based categorizer – From mail recipient, solution will understand whether the recipients belong to Enterprise Applications, Desktop, Telecom, etc. and their related products like Sharepoint, Samsung, etc.
    • Xpresso based categorizer – Xpresso will categorize the emails in Incident Resolution, Service Requests, Questions, Training, etc by analyzing the raw email text body
  • These Combined Categorized output is put into Email Categorization Database, which can also be accessed by communication expert to view that, send alert message to concerned stakeholders and reply to the senders from the solution itself
  • These categorized outputs will also trigger Auto response to the senders, based on certain predefined business rules, specified by client team

Business Benefits

This solution was designed to increase the operational efficiency of the communication expert, who earlier used to read each and every email manually and categorize them.

It helped to reduce the time of the overall survey process by 80 Man Days.

Supplier satisfaction has also increased from 63% to 84% after implementing the solution.