Course teams assemble resource packs that students do not open, and the reason is rarely the quality of the pack.
A well built set of materials on a topic takes several hours to select. It contains the article explaining the concept properly, the primary source, the short film covering what the textbook rushes, and two items included because they disagree with each other. Issued as a link, it produces a predictable distribution: a small number of students work through it, a larger number open the first item, and most open nothing.
The common diagnosis is motivation. A more useful diagnosis is that a list of twelve items communicates nothing about how they relate, and a student facing it has to work out where to start before doing any of the actual work.
The pack cannot say what it is for
A resource pack carries selection and loses reasoning, which is the asymmetry affecting reading lists generally.
The lecturer knows item three is the one to read first, that items six and seven contradict each other deliberately, that item nine is included because its argument is wrong in an instructive way, and that item eleven is optional unless a student is going further. None of that is visible. Annotations help and are rarely read, because a student not reading the sources is not reading the notes about the sources either.
What is missing is orientation, and orientation is spoken rather than written in most teaching contexts. It happens in the two minutes at the start of a session, which is precisely the material that does not survive into asynchronous and distance study.
What should be narrated and what must stay a link
This is the decision the format forces, and getting it wrong produces something worse than the pack.
Narration suits the connective material: why these items, what order, what to look for in each, where they disagree, what the student should be able to do afterwards. This is the two minute introduction, made durable.
Narration is a poor substitute for the sources themselves. Summarising a primary source into narration removes the thing that made it a primary source, which is the student’s encounter with the original wording. A generated summary of a historical document is a secondary source, and a cohort that only meets the summary has not done the work.
Three rules cover most cases. Anything the student is expected to interpret stays a link. Anything the student is expected to quote stays a link. Anything the student needs to return to during an assessment stays a link.
What is left, the orientation and the connections, is usually four to eight minutes.
What conversion involves in practice
Platforms that create educational videos with AI take documents as input and generate an outline, narration and a sequence of scenes. Leadde accepts Word, PDF, PowerPoint and plain text, or pasted text, and offers four levels of detail running from a summary of roughly one to three minutes up to a comprehensive treatment.
For this use the shortest settings are usually correct, and course teams routinely choose longer ones because the source material is long. The source material being long is not a reason for the orientation to be.
Three optional context fields have more effect on the result than any other setting. The audience field determines what the narration assumes: described specifically, for example as a first year student who has met the topic once and does not yet have the vocabulary, it defines terms before using them. The speaker background field sets the position it speaks from. The objective field determines emphasis, and generating with the objective of preparing a student to read the sources produces something quite different from generating with the objective of explaining the topic.
Register is worth setting deliberately. Leadde offers Formal, Neutral, Objective, Structured, Analytical, Explanatory, Narrative and Persuasive, and orientation material is served by Explanatory rather than by anything more assertive.
One capability suits resource packs specifically. Leadde’s knowledge base accepts batches of source files, up to ten at a time, so a set of materials can be uploaded together and drawn on rather than converted individually. For a module whose pack runs to a dozen documents, this is the difference between one preparation task and a dozen.
Preparation time, honestly
The realistic figure for a first attempt is an hour, of which perhaps forty minutes is deciding what belongs in narration and what stays a link. That decision is the pedagogical work and it does not get faster with practice, because it differs for every topic.
The generation itself takes minutes. Reading the generated script before use is not optional and takes about as long as the video runs, since automatic summarisation removes qualifying conditions before anything else, and in teaching material the condition is frequently the thing being taught.
Reuse is where the time is recovered. The same orientation serves the same module next year, amended where the pack changed, which means the second year costs the forty minutes only where something moved.
Keeping the sources reachable
The video introduces a pack and must not become a substitute for it, which is a design problem rather than a good intention.
Three practices help. The video states at the start that it is an introduction and that the sources are the assignment. Item references in narration match the labels in the pack exactly, so a student hearing about the third item can find the third item. And the pack remains the assigned artefact, with the video attached to it, rather than the other way round.
Attribution deserves attention where materials are openly licensed. Openly licensed resources carry attribution conditions that a generated summary can silently drop, and the framework set out by UNESCO on open educational resources is explicit about what reuse requires. A narrated introduction describing an openly licensed source should name it.
Access, and the students this matters most for
Captions belong on the first version rather than being added later. A meaningful share of student viewing happens on a phone in a shared space with the sound off, and captions for prerecorded video are a WCAG success criterion.
Where a cohort includes students working in another language at home, a finished video can be translated with the narration and on screen text handled together, producing a parallel version rather than replacing the original. For orientation material specifically this is more valuable than translating the sources, since the sources are usually the thing the student is expected to work at in the target language.
Three kinds of material that should not be narrated
Reference material should not, including glossaries, timelines and formula sheets. These are consulted rather than watched.
Anything requiring the student to work at their own pace should not, including worked examples where the point is stopping to attempt the next line.
Anything the pack includes because it is difficult should not. The difficulty is the assignment, and narrating past it removes the reason the item was chosen.
Try it on the pack students already ignore
The best first candidate is an existing set with poor engagement on a topic taught more than once, so the preparation cost is amortised.
If more students open the sources, the orientation was the missing piece. If they still do not, the problem is the pack or the assignment, and a video was never going to be the fix.
