The key difference between testing a small and a larger group of students is typically the impossibility to use an e-assessment interface for a very large group of students in one room, especially in mathematics courses. Traditional testing causes overloading of the teacher when evaluating students´ results, and delay in releasing of the results. Moreover, the detailed results and test content are not directly stored in the database for further evaluation, which is a condition for optimization of learning environment. To overcome these problems, new teaching model along with method of assessment with enhanced software support was designed. Experimental data collected during three experimental academic years were compared to traditional assessment data to evaluate the efficiency of the proposed solutions. Results have proved that the overall productivity was significantly increased – more than eleven times without compromising quality. Internal dependencies among parameters of the automated assessment procedure were identified, analyzed and used to maximizing the impact of automation, as well.