The blank page is the most expensive part

Part of a single argument: the engagement gap is a workload problem.

The distribution of preparation effort is not even across the task. A disproportionate share sits in the first twenty minutes, before there is anything to react to.

That is also where most abandonment happens. The sessions that never got their interactive element usually did not lose it at the end, in a considered trade-off. They lost it at the beginning, when someone opened a blank page, felt the size of the task, and decided to do it next time.

Reacting is cheaper than generating

People are dramatically faster at judging than at producing. Given a draft, a competent person will improve it in minutes and enjoy doing so. Given a blank page and the same brief, the same person may take an hour and find it unpleasant.

This is not a difference in skill or motivation. It is the cost of the first move, and it is large enough that it deserves specific attention rather than the general encouragement it usually gets.

It also explains a common pattern: the person who says they are bad at writing questions is often excellent at fixing bad ones. They are not bad at the task. They are bad at starting it, which is a different problem with a different solution.

Anything is better than nothing

The starting draft does not need to be good. A rough outline, last quarter's version, a first pass generated from the source document, a colleague's version of something adjacent. All of them convert the task from generation into editing, which is where people are fast.

Teams resist this because a mediocre draft feels like it will drag the final result down toward its own quality. In practice the editing pass corrects it, and the relevant comparison is not between a mediocre draft and a great original. It is between a mediocre draft that becomes a decent session and a blank page that becomes nothing.

Where a first pass comes from

It almost always already exists. A policy document, a deck from a previous quarter, a recording of the last time someone covered this, a help centre article, a long email that someone wrote to explain the same thing to one person.

Turning any of those into a first draft is mechanical work. Mechanical work is exactly what should not be consuming an expert's afternoon, and it is the clearest case in this whole area for handing a step to software: nobody's judgement is being replaced, because no judgement was involved in the retyping.

Keep the judgement where it belongs

The draft is not the deliverable and should not be treated as one. What matters is where the expert's time goes.

It should go to deciding what actually matters, spotting what is wrong, and knowing what this particular audience will misunderstand. That is irreducible and valuable. The part that was always going to be transcription and formatting is neither.

A good test: if the person reviewing the draft is mostly fixing facts and emphasis, the process is working. If they are mostly reformatting, it is not, and the draft was produced in the wrong shape.

The failure mode to watch for

A draft that is too polished is worse than a rough one. It invites acceptance rather than editing, and it hides the places where it is confidently wrong.

If you are generating first passes, it is worth keeping them visibly provisional. Rough drafts get edited. Finished-looking drafts get published, including the parts nobody checked.