<p>Student agency is increasingly recognized as a critical component of effective, equitable learning, yet it remains underemphasized in the design of personalized and adaptive learning (PAL) technologies. While PAL systems aim to tailor learning experiences, they often do so through automated mechanisms that exclude learners from meaningful participation in their own educational trajectories. This disconnect is especially pronounced in low- and middle-income countries (LMICs), where PAL tools are expanding rapidly but frequently lack student-centered features. To address this gap, we introduce the Learner Adaptive Domain Agency (LADA) model, a conceptual framework that integrates student agency into PAL system design by focusing on autonomy, competence, motivation, and connectedness. Drawing on Social Cognitive Theory and Self-Determination Theory, we develop a 29-item rubric (AiPAL) that translates these theoretical principles into observable design criteria. We apply this rubric to 40 PAL systems deployed across 70 countries to evaluate how agency-supportive features are operationalized in practice. Results reveal persistent limitations in how systems support student goal-setting, inclusive access, co-regulated learning, and reflective feedback. Fewer than half of the platforms enabled shared control over learning pathways, and most relied on extrinsic motivation strategies rather than intrinsic supports. The findings highlight both the urgency and feasibility of reorienting PAL design toward agency-centered models. We conclude with actionable recommendations for designers, educators, and policymakers to advance equitable, empowering digital learning systems.</p>