Plan an AI Product Video with a Three-Shot Storyboard
Plan a 12-second AI product video with a three-shot storyboard, separate reference images, copy-ready prompts and a practical product-checking shot list.
A useful AI product video storyboard gives each clip one job. For a mug with a large handle, the opening can establish the shape, the middle can show a hand using it, and the closing can leave room for a short title.
This is a production plan for a 12-second edit. The three original stills below define the intended shots; no finished video is being presented. We checked the site's video controls and Adobe's video-planning guidance on 3 November 2026 before choosing the inputs and actions.
Turn one product feature into a visible action
The brief uses a fictional dusty-blue ceramic mug. Its large D-shaped handle is the feature to show. There is no capacity, heat-retention or comfort claim to prove with invented footage.
| Part of the brief | Decision for this plan |
|---|---|
| Product identity | Straight blue body, cream interior, rounded D handle on the viewer's right |
| Main message | Show the handle clearly, then show an ordinary grasp |
| Setting | Pale oak desk, cream notebook, soft daylight from the left |
| Final delivery | A landscape 12-second edit made from three separate clips |
| Text | Add the closing line in an external editor after checking the footage |
For a real product, replace these drawings with photographs you own or can use. A generated cup with a larger handle would misrepresent a product that has a smaller one. Establish the actual silhouette before asking it to move. The product-photo guide covers preparing that starting image.
A three shot product video plan for a 12-second edit
The timings describe the final edit. They do not ask a video model to make a complete advertisement in one generation.
| Time | Shot, action and acceptance |
|---|---|
| 0 to 3 s | Establish. Slowly push toward the resting mug. Keep the whole cup and handle opening visible. Reject a handle hidden behind the body. |
| 3 to 8 s | Detail. The existing grip raises the mug slightly, then holds. Keep a readable grip and stable cup. Reject fingers crossing the ceramic or a bending handle. |
| 8 to 12 s | Close. Hold a wider view of the mug resting on the desk. Keep clear space above and to the right. Reject new objects or generated lettering in that space. |
The second shot carries the demonstration, so it gets five seconds. The opening needs less time. The closing gives the viewer a stable picture while the later title appears. These are editorial choices for this brief, not a platform requirement.
Establish the whole object
Start with the mug resting on the desk. The handle needs a light background behind its opening. A busy object in that gap would weaken the feature before motion even begins.
Use the matching still as the image input, then give a restrained motion instruction.
Make a single continuous shot from this image. The mug stays still on the desk while the camera slowly pushes closer. Keep the blue ceramic body, cream interior and right-side D-shaped handle unchanged. Preserve the left-side daylight. Keep the entire handle visible. No cuts, new objects or lettering.
Give the hand one action
The detail input already includes the grip. That makes the starting condition explicit. Asking a hand to appear, pick up the cup, rotate it, pour a drink and set it down would introduce several separate events into one short clip.
Keep the camera fixed on this grip. The hand raises the mug a little, then holds it steady. Preserve the cup's cylindrical body, cream rim and D-shaped handle. Keep the fingers attached to the same hand and the handle rigid. No drinking, pouring, rotation or additional hands. No text.
Review the whole action before accepting it. A convincing first frame does not establish that the handle remains intact halfway through the lift.
Leave the closing frame quiet
The wider closing composition creates space for a later title. The cut from the hand detail to the resting cup is intentional; this plan does not promise a seamless handoff between generations.
Hold this composition as one continuous shot. Keep the mug resting on the desk and the camera locked. Preserve the blue glaze, cream interior, handle position and soft daylight. Keep the wall above and to the right uncluttered for a title added later. No hand, steam, moving props or generated text.
Prepare a separate input for every shot
Open the AI Video Generator. Choose a model that accepts image input, attach the still for the current shot, and enter that shot's motion instruction. Check which attachment is selected before each request. A message saying “use the detail shot” is less reliable than supplying the correct detail image.
On the date checked, MiniMax H3 offered image input, a 5-second duration, and a 16:9 option. Five seconds per source clip would supply enough footage for the planned 3-, 5- and 4-second cuts, provided each generation contains usable motion. A model's duration setting does not guarantee that every frame will pass review. Recheck the current options and maximum credits after changing the model or settings.
Use the same aspect ratio for all three clips. The source stills here are landscape compositions; inspect the framing if the video model fits them into a different shape. A vertical version deserves its own planned composition. The aspect-ratio and safe-area guide explains that decision without stretching the cup.
The preparation folder should contain the three named stills, the three motion prompts, a product photograph for comparison, and the approved closing copy. Keep brand artwork and any permitted music separately. You can download the AI product video shot list and brief with prompts to adapt them.
Check the product before assembling the clips
Compare each result with the same product reference. Inspect the beginning, middle and end, then watch the movement continuously. For this cup, check the handle's opening, where it joins the body, the cream interior and the body's proportions. A good-looking change can still be a product error.
Check the hand separately. Look for merged fingertips, fingers crossing the ceramic, a floating grip or a shadow that stops following the cup. Reject the shot if the feature being demonstrated becomes unclear. Keep the strongest usable version of each shot rather than repeatedly changing every prompt at once.
For packaging, read every visible word and inspect the logo throughout the motion. A static logo overlay will not automatically track a turning package. If accurate packaging cannot survive the shot, simplify the movement or film that detail conventionally. The camera-movement guide helps reduce a shot to one understandable move.
Finish the edit and titles outside the generator
Download the accepted clips and assemble them in an external video editor such as Shotcut. Put them in storyboard order, trim to 3, 5 and 4 seconds, and check the 12-second sequence before adding text. If a clip has too little usable footage, revise the plan or regenerate that shot; freezing or slowing damaged frames will not repair the cup.
Add a closing line such as “Meet the wide-handle mug.” as a text layer or text clip in the external editor. Shotcut documents both text filters and separate transparent text clips. Keep important copy inside the intended placement's safe area. Export the actual video file as well as saving the editable project.
MiniMax H3's checked interface showed model-defined sound with no sound switch. Review the downloaded audio and handle muting, replacement or music in the external editor. Do not assume the clip will arrive silent.
Before publishing, watch the exported file on a phone. Read the title, inspect the two cuts, and check that the product still matches its reference after export. Keep the stills, approved clips and project together so a later copy change does not require rebuilding the whole plan.
