I once reviewed a course that was immaculate. Clean modules, consistent naming, every file in its place. It looked designed. Learners moved through it the way you move through an airport — processed, on time, and taught almost nothing they kept.
An LMS is a container. It stores, sequences, and tracks. It cannot decide what deserves attention, where a learner should struggle productively, or when a human should step in. Organised is not designed, and a well-organised content dump is still a content dump.
What the container cannot do
The platform gives you shelves. The design is what you put on them and in what order and why. Selwyn’s caution about ed-tech holds here: the tool tends to get credited with the learning it merely houses (Selwyn, 2016). Reich’s history of large-scale platforms tells the same story — distribution scaled, learning did not follow automatically (Reich, 2020).
An LMS is not a learning experience. It is a container. The design still has to do the human work.
Interaction is a decision, not a click
The container will happily register activity that is not learning: pages turned, videos autoplayed, boxes ticked. In my own courses, the interactions that worked were the ones that made a learner decide something — choose, commit, and then see the consequence — not the ones that only asked them to press Next. Building one real decision into a single page did more than a dozen embedded videos. And every hop out of the container to a video here and a document there is a cost, paid in a learner’s attention.
Why AI raises the stakes
AI makes it trivial to fill a container fast, which makes the tidy-but-dead course more common and more convincing. The shell looks fuller than ever. The question the shell cannot answer is the only one that matters: inside all this, where does a person actually change?
The move
Audit a course by walking it as a learner and marking every moment you are asked to make a real decision. If the decisions are sparse, the container is doing the talking and the design has gone quiet.