The learning factory approach has proven valuable in education, lifelong learning and research over the past decade. Learning factories have well-defined goals and enhance participants’ knowledge and experience with specific technologies. However, existing research often lacks systematic conclusions on how the design of scenarios affects learning outcomes, thus hindering theory generation. The CIRP Collaborative Working Group’s morphology is an initial attempt to characterise learning factories, but further study is needed to understand the scenario design’s impact. This paper uses AI-supported bibliometric analysis to identify trends in learning modes, target groups, topics and didactics. A review of 421 studies found that most research has focused on transferring closed-case learning outcomes to students and professionals, with a growing trend of knowledge creation in open-ended, interdisciplinary learning factories. These insights conceptualise learning factories as a multimodal concept, requiring distinct design principles based on the learning target and supporting strategies for cost reduction, knowledge extension and innovation.