Reflection remains a defining feature of music-teacher preparation (Conway, 2020), yet preservice teachers vary widely in how they understand and value reflective writing (Conway, 2001; Hourigan & Schiebe, 2009; Kruse, 2012). This pilot study examines whether undergraduate music-education students perceive using large language models as relevant to reflective practice when it is made available as a resource for music teaching practicum experiences.Students enrolled in a secondary instrumental methods course were required to establish a digital notebook with AI-assisted features that operate exclusively on user-uploaded materials rather than general training data. They were encouraged to upload lesson plans, instructor and cooperating-teacher feedback, and daily reflections, and to use the notebook’s pattern-identification tools to analyze their own materials. Across the semester, students completed five targeted reflections responding to instructor-provided questions. The instructor added individualized comments to each entry to model dialogic feedback and deepen reflection. The use of reflective practice was embedded throughout the course in class and practicum activities (Schön, 1983). The use of a digital notebook provided ample opportunities for students to draw upon an LLM (Large Language Model) as they saw fit.An anonymous end-of-semester survey was used to collect (a) baseline conceptions of the purpose and value of reflective journaling, (b) patterns of LLM feature use, (c) reasons for adoption or non-adoption, and (d) perceptions of instructor feedback relative to LLM-generated insights. Likert-scale and open-ended items were used to analyze to identify engagement profiles such as (a) students who found LLM helpful for pattern detection, (b) those who resisted it as inauthentic, (c) those satisfied with traditional reflection, and (d) those who adopted organizational functions while avoiding interpretive ones.The project explored a foundational question of LLM integration: Under what conditions do preservice music teachers perceive models as compatible with reflective practice? Decisions to engage with AI features were interpreted as affirmations of the compatibility of model affordance and students’ conceptions of reflection. Reflective orientations and corresponding AI engagement patterns are presented, along with evidence-based recommendations for integrating LLMs in ways that support authentic, dialogic professional learning and respond to both preservice teachers’ optimistic engagement and resistance to AI.