Abstract
This conceptual article proposes the Learner’s Discerning Practice Model for working students in MBA and specialized master’s programs in East and Southeast Asia. These students study while employed, and generative AI can most easily do the reflective work that connects course concepts to professional decisions. Recent studies report declining decisiveness and self-reflectiveness among graduate students who write with generative tools, and critics observe that assessed, formulaic reflection was automatable before such tools. This article argues that reflective practice resists automation when it is private, rhythmed, time-limited, and answerable to action: each reflection commits the student to a decision before that decision is made, and is judged by what the decision produces. From the reflective practice tradition, self-regulated learning theory, and four classical learning theories, the article derives five practices (Experience, Synthesize, Apply, Frame, and Grow), each producing a dated artifact on a fixed cadence, across four phases of the graduate journey (Pre-program, Coursework, Thesis, and cross-cutting Resonance between study and work). No new empirical data are reported. Each practice has known antecedents; the contribution is their integration into an architecture a working student can sustain. The article positions the model against recent work in the reflective practice literature, states its limitations, and proposes a test design.