An Intelligent Tutoring System (ITS) is software that dynamically modifies the presentation of educational material to students based on Artificial Intelligence. ITSs enable tutoring to be done by computers; teaching is customized for the student using the system, enabling faster learning; and there is a potential to enable scaling up, teaching of millions of students, using online platforms or platform independent programs which is not possible in the case of human tutoring. In this paper we focus on the cost implications of an ITS designed for teaching Discrete Mathematics at the undergraduate level at the University Of Botswana. The ITS discussed, DS-ITS, makes use of probabilistic graphical networks to relate core concepts of Discrete Mathematics concepts with verifiable outcomes (demonstrable knowledge). A student's understanding of a concept is inferred; it can be viewed as the cause of failing to answer a question correctly. Our contribution is in estimating the cost required to build a prototype, which can be used for certain Discrete Mathematics topics as taught at the undergraduate level. We compare the cost implications with the alternative of not having such a tool and instead relying purely on lecturers and teaching assistants. Our findings show the benefit of developing such a tool.