The use of social robots in clinical settings for stuttering therapy of preschoolers by Speech-Language Pathologists (SLPs), remains a relatively unexplored area. To enable Human-Robot Interaction (HRI) researchers to study this domain in the future, it is crucial first to identify the potential applications of social robots in clinical settings and understand the relevant considerations for these interactions. To explore this, we conducted interviews and shadowing sessions with SLPs working in stuttering clinics. Our data collection and analysis adapted a Charmaz/Constructivist Grounded Theory Methodology, ensuring our findings were based on the collected data. The results suggest practical applications along with interaction and communication design considerations for integrating social robots into speech therapy sessions for preschoolers who stutter. Key applications for social robots include engaging children through various games and providing a consistent speech model, among others. Essential properties of interaction design for sessions include being interactive, structured, and supervised by SLPs. Additionally, the robot’s communication could involve slow, relaxed, personalized speech, and other techniques to support children’s speech goals.