LEA's course-agnostic architecture was operationalised across three courses with no changes to the orchestrator, mastery tracker, or scaffolding engine¶

Claim · All claims
Evidence · 1 study · design r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? Deployment across three courses spanning two disciplines and two academic levels required only course-specific RAG and KC artifacts, with no changes to the orchestrator, mastery tracker, or scaffolding engine. → Rumble 2026

Evidence¶

Rumble 2026¶

Rumble, T., Zarrin, J., Lovell, P. G., Falconer, R. (2026). Learning Engagement Assistant (LEA): Cross-Course Scalability and Classroom Evaluation of an Agentic AI Tutoring System. arXiv. https://arxiv.org/abs/2607.13370 doi:10.48550/arxiv.2607.13370

q2 · i? · design · r2

Deployment methodology across CMP511 (MSc computing), CMP202 (UG computing), and PSY555 (MSc social science), each onboarded through the identical three-stage pipeline. The authors summarise the design claim as "the orchestration layer generalises; the knowledge layer adapts."

"This design was tested through deployment across three courses spanning two disciplines and two academic levels (Table 3); each required only course-specific RAG and KC artifacts, with no changes to the orchestrator, mastery tracker, or scaffolding engine."

Discussion¶

Related Claims¶

• Quiz mode's course-blind Python question generation was the only cross-course failure requiring a code change during onboarding — reports the opposite
• Over two semesters the LearnAI deployment reached 18 courses (293 enrolled students) and 35 co-creation clients who produced 36 portfolio websites and over 20 deployed applications — related
• RAGAS Faithfulness declines with curriculum distance from LEA's original course (0.685 to 0.499), while Answer Relevancy and Context Precision remain broadly stable across courses — a narrower finding that bears on this claim
• In a graduate QIS course pilot, the specialized Spoke-and-Wheel architecture showed no task-boundary failures across 75 student interactions, though no controlled comparison against the prototype was run — related
• Many agentic multi-agent implementations remain at prototype or conceptual stage with limited large-scale deployment and empirical validation — related
• Class size, course level, and discipline did not explain significant variation in how achievement gaps changed under active learning — related