A Call Center Simulation Study: Comparing the Reliability of Cross-Trained Agents to Specialized Agents

Call centers are an important function of most companies’ day to day business activities. They are often the link between a company and its customers and hugely impact the customer’s perspective or point of view (POV) of a company. A call center in the most general sense is a place, representing a business, which receives inbound calls from customers and/or makes outbound calls to customers, the latter being most commonly referred to as telemarketing. There was a time when a typical call center strictly consisted of agents who handled inbound/outbound calls; these agents are considered specialized agents. Generally speaking, a specialized agent is one trained, in-depth, in a particular area of knowledge.

Most businesses have transgressed from your typical call center into contact centers. Contact centers operate essentially the same as a call center but interact with the customer in a variety of ways including, but not limited to: Phone, Mail, Fax, Email, and Internet (via online chat and instant messaging applications). The dynamics of these kinds of call centers has caused an increase in the need for agents to become more diverse in their talents and abilities to handle different types of calls. This has lead to specialized agents becoming general or “cross-trained” agents in which they are trained, broadly, over several areas of knowledge.

The purpose of this thesis is to compare specialized agents to cross-trained agents and through the use of simulation, determine which of the two are more efficient and reliable in their ability to service the customer. This thesis has three major components: Simulation, Reliability Analysis, and Comparison. The results indicate that a cross-trained model is more reliable and efficient than a specialized model. Performance metrics common to call center literature, simulation, and Lean reliability systems were used to determine the effectiveness and reliability of the two models.