We aim to improve the communication process of a teacher with students during lectures using question answering. Our work is focused on the analysis of students’ answers to support the teacher in his or her lecturing. We work with students’ answers to open questions, where it is impossible to identify finite number of solutions. In large classes it is impossible to react in real time to such answers since their evaluation is time consuming. We propose our own approach that helps the teacher by grouping similar answers. These groups are created based on proposed method employing text classification and clustering. Proposed method automatically estimates a number of clusters in answers using combination of k-Nearest Neighbors (KNN) algorithm and affinity propagation. We evaluated the method on real data in Slovak language collected from the course Principles of Software Engineering using real time presentation system ASQ.