Online classes are increasingly gaining popularity in both children's and adult education, although, the challenge of the students’ engagement remains. Evidence has shown that co-teaching is beneficial for both tutors and students, however, especially on a university level, tutors are not always able to co-teach classes. Additionally, socially assistive robots (SAR) have been proven advantageous in educational roles, as they deliver teaching material and provide support to students across various educational levels. This research paper specifically examines the performance of SAR in place of university co-tutors in online classes, in the field of computer science, measuring three outcomes, knowledge acquisition (academic test), enjoyment level (questionnaire), and engagement (qualitative and quantitative analysis of verbal and chat discussions and threads during class). We also aim to fill in the gap of having structured human-robot collaboration guidelines for educational environments. This will be accomplished through a sequence of three experiments: 1. A co-tutored lecture involving one human and one robot, 2. A co-tutored lecture involving two humans, 3. A second instance of a co-tutored lecture involving both human and robot. Results indicated that students taught by a collaborative team of human and robot tutors exhibited a greater level of knowledge acquisition, while maintaining comparable levels of enjoyment, in comparison to those instructed by a collaborative team of two human tutors. Moreover, engagement remained high during the lectures, and we are addressing the differences in students’ communication channels and discussed topics when lectured by two humans vs a human and a robot. Finally, the human-robot collaboration managed to eliminate the negative outcomes of the presence of a robot, such as the surprise effect based on the literature the students face the first time they are being lectured by a robot-tutor.