A system, method and computer program product are provided for monitoring voice application calls over a network. Initially, a first leg of a voice application call is identified. Next, a unique key is created based on the first leg of the voice application call utilizing a first session object. Further, a second leg of the voice application call is identified. Another unique key is created based on the second leg of the voice application call utilizing a second session object. In use, both the first and second session objects are associated with a single application object.