Changing the traditional approach to experimentation in science teaching poses a significant challenge. This research discusses the characterization of physics experiments, not solely based on the topics they cover, or the materials used in their setups, but rather in terms of the opportunities they offer to enhance classroom interaction and, consequently, contribute to the development of scientific thinking skills. This study is an action-research intervention in the classroom, involving four groups of undergraduate physics students from a public university in Bogotá, Colombia. This process allowed us to categorize experiment typologies, including discrepant, homemade, illustrative, research-based, recreational, crucial, mental, and virtual experiments. The most relevant data were derived from students’ productions and researchers’ reports after each class. Through content analysis techniques, we were able to categorize the information and derive insights into the richness that each experiment typology brings to classroom interactions. The results reveal shifts in the perception of the role of new teachers, who are no longer seen as mere possessors and transmitters of truth. They now comprehend strategies for effectively engaging students as the class progresses, fostering the exchange of ideas, reflections, debates, and questions among students themselves, their peers, and the teacher.