Translate the picture in your head
You might imagine a quiet opening, one satisfying action, and a clean finish. Put those three things into words. “Start on the empty table, show the cup being placed, finish with steam rising” is already a workable direction.
The trouble starts when the imagined video includes shots you never recorded. Before uploading, check that the opening, action, and ending exist. If they do not, change the concept or film the missing piece. More adjectives will not solve a missing shot.
Let the footage do some of the work
You do not have to prescribe every cut. Tell Shorty what matters and what is flexible. Perhaps the pouring action must remain intact, but the order of the detail shots can change. Clear priorities leave room for the assembly to work.
Add a reference when the rhythm is difficult to describe. Explain which part you mean: the pauses, the order of the reveals, or the amount of text. Use reference material and audio you are allowed to use, and check the finished soundtrack.
Expect a conversation
The first output makes your idea visible. Watch for what feels right before listing what feels wrong. Then request a focused revision, such as a slower reveal or a different closing shot.
This workflow is useful when you want to direct a result without arranging every clip by hand. It still needs your eye. Check that the edit shows the right product, preserves the meaning of speech, and does not imply something your footage cannot support. Download the version you have actually reviewed.
A brief to adapt to your footage
Example direction. Replace the subject, shots, and wording with your own.
Make a calm 15-second coffee video from these clips. Start with the empty cup, show the pour without interrupting the main action, and end on the steam close-up. Use few cuts and no sales headline. Keep the natural sounds if they are clear.