Create an AI video by separating story, visuals and motion.
Do not start with “make a 30-second video.” Build a short shot plan first, then generate only the assets and motion you actually need.
Best-suited tools
RunwayStrong fit for generative video, image-to-video and editing workflows.Adobe FireflyUseful when AI video belongs inside a broader Adobe creative workflow.ChatGPTUseful for concept, shot planning, prompt iteration and supporting image creation.MidjourneyUseful for creating strong visual keyframes before animating them.
Workflow
01Write the objective
Platform, audience, product/message, duration and one action you want the viewer to take.
02Create a shot list
Describe 4–8 shots with subject, camera, motion and duration.
03Generate keyframes
Create consistent still images for important scenes before asking for motion.
04Add motion
Use concise motion prompts: camera move, subject action, environmental motion and timing.
05Edit and verify
Check continuity, product fidelity, text, timing, audio rights and CTA clarity.
Shot prompt template
SHOT [NUMBER] — [DURATION]
SUBJECT:
[WHO / WHAT]
SCENE:
[ENVIRONMENT]
CAMERA:
[SHOT SIZE + CAMERA MOVEMENT]
ACTION:
[ONE CLEAR SUBJECT ACTION]
ENVIRONMENTAL MOTION:
[WIND, LIGHT, PARTICLES, BACKGROUND MOVEMENT]
LOOK:
[LIGHTING, COLOR, MOOD]
CONSTRAINTS:
- preserve product/character identity,
- no unrequested text or logos,
- believable motion and anatomy,
- avoid abrupt camera jumps,
- keep continuity with the previous shot.
Verification checklistWatch once with sound off for continuity and composition, once focusing only on product/character consistency, and once focusing only on timing, text and CTA.