The field notes · October 4, 2026
A simple workflow for vertical AI videos
A repeatable process to plan, generate, review, and edit a short vertical video.
Write a one-sentence brief
Decide who the clip is for and what one thing they should see. An opening product reveal, a mood shot, and a dance study require different framing. Pick the goal before choosing effects.
Build a shot instead of a whole film
Use a short clip with a single action. Set the aspect ratio in your generation tool, and keep important details away from edges that may be cropped or covered by captions.
Generate and compare
Save your prompt and settings. Create a small number of alternatives and compare continuity, framing, and timing. Choose the clearest version instead of adding more motion to fix a weak result.
Edit with restraint
Trim dead time, use readable captions when needed, and preview on a phone. Use music, voices, and images you have rights to use. Avoid adding claims that the visuals cannot substantiate.
Export and learn
Check the final file for visual glitches, audio levels, and unwanted lettering. Export according to your publishing platform’s current requirements. Record what worked so your next clip starts with evidence rather than a blank prompt.