How to Make AI Product Videos from Photos: A 3-Shot Plan

Write a one-sentence brief before generating: “Show [product] to [viewer], make [one visible feature] clear, then invite [next action].” A focused brief helps you reject attractive footage that says little about the item.

For an insulated bottle, a useful first brief is: “Show a commuter the bottle’s compact shape and textured grip, then invite them to choose a color.” A video of the bottle floating through a fantasy landscape might look interesting, but it would not necessarily answer those questions.

Separate appearance from performance claims

A close view can show the intended finish. It cannot prove leak resistance or temperature retention. Use actual product evidence for performance claims, and do not let generated scenes imply a test you have not performed.

The examples here are planning templates. They have not been run as a measured campaign, and they carry no promise of ad performance.

Give each shot one job. You can aim for roughly two or three seconds per shot in the final edit, but generate using the durations available in your chosen model and trim the approved clips afterward. The edit plan is not a claim about exact generation timing.

Shot 1: show the whole product

Start with a clear view that makes the item recognizable. Let a modest camera move bring it into focus as the main subject. Avoid beginning so close that viewers cannot tell what they are seeing.

Shot 2: show one decision-making detail

Choose a real feature visible in your source: a zipper pull, textured grip or cap shape. Use a separate detail photo when the original does not show it clearly. Do not ask the model to invent hidden construction.

Shot 3: leave space for the next step

Finish on a readable product view with space for text. Add the offer, website or call to action later as an editable overlay. This lets you correct the wording without regenerating footage.

Use photos of the exact item and variation you intend to advertise. Check its outline, color, attachments and printed information at full size. An attractive reference of a similar item is not enough for an accurate product ad.

• Opening photo: the entire item is visible, with clear separation from the background.
• Detail photo: the feature from your brief is in focus and large enough to inspect.
• Closing photo: the composition leaves space for your planned text.

All three photos should agree on color and configuration. If only one suitable photo exists, make one restrained shot first. Reusing an accurate still in the edit is preferable to inventing a second angle.

Choose the delivery shape early

Decide whether your intended placement needs a tall or wide composition. Inspect the source crop before generation and choose a supported aspect ratio in the tool. Keep the product and planned text away from the edges; check the destination’s current overlay and crop requirements before export.

These prompts use the illustrated bottle as a planning example. Replace it with the visible product in your photo, and remove any detail your image does not support. For help separating subject and camera motion, read the image-to-video prompting guide.

Opening prompt

The bottle stands on the pedestal. The camera makes a small, slow move closer while keeping the entire bottle in frame. The shot settles into a clear front view.

Check whether the bottle remains the same width and the base stays attached to the surface.

Detail prompt

The camera holds a close view of the bottle cap and upper textured surface. A soft highlight moves gently across the surface while the bottle remains still.

Use a detail source image. A lighting change is an aesthetic direction, not evidence of a particular material or coating.

Closing prompt

The camera holds a steady medium view of the upright bottle. The product remains centered below the open space in the upper part of the frame. The composition stays calm through the end.

Add your own text overlay in the editor. Inspect the blank area too: an unexpected object behind the text can make the closing frame unusable.

The product ads workspace provides a place to explore product-video concepts. For direct control over a photo and its movement, use the image-to-video generator.

1. Upload the source photo for your first shot. If you use inspiration from the product ads workspace, replace its subject and scene details with your own brief.
2. Enter one of your adapted motion prompts. Confirm that the selected model supports the image workflow you need.
3. Review the displayed duration, aspect ratio and credit cost before submitting. Make one baseline clip first.
4. Inspect the result against the source and record what to change. Keep a clip only if the product remains suitable for the intended use.
5. Repeat for the detail and closing shots using their corresponding photos. Assemble the approved material in your video editor.

Different models offer different controls. This workflow does not require a particular subscription, fixed generation cost or guaranteed output resolution; check the current options in the workspace.

Watch the clip at normal speed, then inspect the start, middle, end and any frames with a turn or close-up. A convincing first frame is not enough.

Keep a simple review sheet: shot name, source photo, prompt, selected settings, visible defect and approval decision. Change one variable in the next attempt so you can learn from the result.

Trim each approved clip around its clearest moment, then place the shots in opening–detail–closing order. Cut before the viewer has to wait for the next useful piece of information.

• Add a short, accurate opening line that matches the product brief.
• Use readable text overlays for names, prices and the next action rather than relying on generated lettering.
• Choose audio you have permission to use and preview the edit both with and without sound.
• Inspect the final crop on a phone. Confirm that the product, text and call to action are visible together.

To compare two versions, change a single element such as the opening line or first shot. Look at the actual viewing and conversion data for your placement; an AI-generated video is not automatically more effective than a still image.

Can I make a product video from just one photo?

Yes, you can start with one restrained shot. Do not assume a single front view supplies accurate information about the back, underside or moving parts.

Can the video show a product working?

It can depict an action, but that is not proof of the product’s real behavior. Use real demonstration footage when an action or result is essential to the claim.

Should I generate the whole ad in one request?

For this workflow, separate shots make review and revision easier. You can replace one weak clip without rebuilding the entire sequence.

What if the product keeps changing?

Reduce the camera angle change and simplify the scene. If the product still cannot be represented accurately, use an approved still or real footage for that part of the edit.