Client Context

IRONSCALES is a cybersecurity company specializing in email phishing solutions. They operate a customer support division which handles and resolves client complaints through a ticketing system. Clients initiate support requests that are routed to a team of agents for resolution.

Project Objective

The main objective of the project was to enhance the efficiency of IRONSCALES' customer support operations. The focus was on developing a scalable system to handle increasing ticket volumes due to rapid company growth, thereby avoiding potential agent burnout and compromised service quality​​. This was achieved by reducing the time to resolution (TTR), agent work time, and agent utilization of the system. 

Design Strategy

Constraints: The project had to consider the synchronization of two office locations in different time zones, training time for a Natural Language Processing (NLP) model, current workflow of the customer support team, integration with their customer support platform, and limitations within their data storage platform.

Main Assumptions: Our team created a simulation to test the effectiveness of our designed system compared to the current system. For this simulation, it was assumed that agents could handle up to three tickets simultaneously and would not receive new assignments if their queue exceeded ten tickets unless all agents had a queue length greater than ten. These assumptions were made based on data analysis and input from the client. The simulation also assumed no status changes, as third-party investigations and client response times were beyond the scope of this project​.


  • Dynamic Ticket Assignment System: This system includes ticket classification using NLP and a ticket assignment model that matches tickets to suitable agents based on historical agent performance, workloads, and locations. The model utilizes an Upper Confidence Bound when measuring agent performance to adjust for uncertainty due to varying levels of agent experience. Additionally, it features a performance monitoring process that automatically alerts managers when a ticket exceeds a typical resolution time threshold.
  • Diagnostic Dashboard: A Tableau dashboard offers a comprehensive view of key metrics and trends in customer support operations, complete with filtering options for targeted analysis. This tool aids in decision-making for customer support management and enables agents to monitor their own performance and progress.

Value and Impact

The value of the Dynamic Ticket Assignment System was shown through quantitative, simulation-based results: a 5% decrease in TTR, a 22% decrease in average service time, a 33% reduction in average queue length, and a 25% decrease in the standard deviation of queue lengths. These improvements enabled faster ticket resolutions, created reduced and balanced individual workload for agents, and overall contributed to the success and efficiency of the company​​. 

The Diagnostic Dashboard provided significant value by saving over 100 labor hours annually. Surveys conducted post-implementation revealed a 100% satisfaction rate among users, affirming its effectiveness and utility in enhancing the operational efficiency of the customer support system​​​​.

Project Information

Fall 2023

Student Team

Max Mobley

Nandan Sanghani

Lasya Akshara

Leya El Fadel

Suchith Munigati

Jordine Jones

Nathanael Mitchell

Faculty Advisor

Faculty Evaluator