The way teachers use language in the classroom has a direct impact on student participation. Expressions that explicitly invite students to ask questions are a powerful—yet understudied—discursive strategy, especially when combined with computational analysis. Addressing this gap, we propose a computational method for detecting these expressions in real classroom settings. As a case study, we analyzed a university mathematics class (on set theory) from an engineering program. The audio was transcribed with OpenAI's Whisper model, and teacher interventions were analyzed using sentence embeddings from the all-MiniLM-L6-v2 model, with manual validation to ensure accuracy. Results indicate these expressions can be precisely identified and exhibit a non-random distribution, concentrated at the session's start and end, suggesting intentional structuring to promote dialogue at strategic moments. To conclude, this study presents a replicable methodology for examining teacher discourse and its connection to active student participation, which is valuable for teacher training and feedback programs.