This paper examines documented CDIO alignment across the full portfolio of Bachelor of Engineering module specifications at DTU. Using AI-assisted qualitative document analysis, 284 publicly available module specifications were analysed with the CDIO Standards 3.0 and Syllabus 3.0 as the formal analytical point of departure. Microsoft Copilot was used to support semantic reading at scale, while interpretive judgement remained with the researcher. The analysis focused on overall documented alignment rather than isolated standards. Each module was categorised on an ordinal scale of documented CDIO alignment: Weak, Weak-Medium, Medium, Medium-Strong, and Strong. Across the portfolio, 8 modules (2.8%) were classified as Weak, 132 (46.5%) as Weak–Medium, 129 (45.4%) as Medium, 10 (3.5%) as Medium–Strong, and 5 (1.8%) as Strong. The findings suggest that strongly documented CDIO alignment is relatively rare in the module specifications, while a large proportion of the portfolio shows some visible alignment without documenting it strongly enough to count as clearly substantial. The overall picture is therefore one of a portfolio dominated by partial or limited alignment, alongside a smaller number of more clearly convincing cases. The study also suggests that stronger documented alignment tends to emerge where several pedagogical elements are articulated in mutually supportive ways within a module, so that the specification presents a coherent learning design rather than a set of separate signals. The paper contributes both a portfolio-level picture of documented CDIO alignment and a methodological discussion of AI-assisted qualitative document analysis in educational quality assurance.
