The Agents of Things (AoT) is a novel concept that unites the physical and cyber worlds and enables things to intelligently interact and communicate with each other. The main goal of this concept is to minimize the constraints on the Internet of Things (IoT), represented by the lack of self-reasoning capability in the things and to correspondingly offer improved applications that can provide value-added services to humans. In this paper, we analyze and classify the main types of software agents that can provide an adequate reasoning capability for the AoT concept. Then, we introduce the Value-added Service Analyzer as a sub-model that produces an optimum reasoning capability for the things in an AoT to provide additional value-added services for humans.