High variability in demand is more prevalent in the service industry than in the manufacturing industry. Additionally, the manufacturing industry is able to rely on inventory to cope with the fluctuation of demand, but customers in service industries must be taken care of promptly when they arrive. To ensure an acceptable service level, the servers must be scheduled according to the changing demand. For simple queuing systems the number of servers can be obtained analytically. However, for many real systems, with changing demand, simulation is a more suitable tool because of its capability to imitate the behavior of the real system. In order to use simulation effectively, the pattern of the arrival rates of customers must be identified first. This can be done using a statistical method called change-point analysis. Change-point analysis is a generalized likelihood ratio test formulated to denote a change-point (a point in time) which is most likely to be the time when the parameter(s) of the distribution of a time-series data changes. This test is done sequentially for every possible segment of the data. Recent research in the area of Statistical Process Control (Hawkins and Qiu, 2003, Hawkins and Zamba, 2005) suggests using limits derived through Monte Carlo simulation to test whether the statistic used is significantly large enough to imply a change-point in data from manufacturing processes. This method is further adopted for Poisson distribution to extend its use in a service environment (Chandra and Nembhard, 2005). In this paper, a case study on a queuing system at BCA bank is studied using simulation in AutoMod software. The model accounts for real situations including customer balking, teller breaks, and changing arrival rates throughout the day. A change point analysis revealed that two types of arrival rate patterns exist for this system. Using this information as input for the simulation, we are able to compare scheduling rules and the corresponding service levels when demand varies. We also investigate the use of administrative employees as temporary tellers during peak hours.