Guide to Creating AI Video Ads from Product Photos

This instructional guide details the process of transforming a simple product photograph into a dynamic video advertisement using an AI-based workflow platform (such as Weavy). The method ensures that crucial branding elements like text and logos remain crisp and unaltered by the generative AI process.

Phase 1: AI Image Generation and Styling

The initial phase involves generating a high-quality, styled product image that will serve as the starting frame for your video ad.

1. Prepare Your Input Assets

  • Product Photo: A clear image of the product you wish to advertise.

  • Style Inspiration Photo: A separate image that dictates the desired aesthetic for your ad, including the environment, lighting setup, and overall mood.

2. Generate Product and Style Descriptions

  • Product Description: Upload your product photo to a Large Language Model (LLM) node and generate a detailed description of the product itself.

  • Style Description: Upload the style inspiration photo to a separate LLM node. Prompt this node to describe the environmental elements, lighting, and colors. Crucially, instruct the LLM not to reference any specific product in this output; it should focus only on the visual style.

3. Concatenate the Final Prompt

  • Use a Prompt Concatenator node to assemble all necessary inputs for the image generation model.

    • Start with a command instruction (e.g., "create a banner of product").

    • Input the Product Description.

    • Input the Style Description.

    • Add a final constraint, most importantly: "no text on the ad". This prevents the generative AI from creating illegible text that would later interfere with your actual branding.

4. Generate the Styled Ad Image

  • Feed the concatenated prompt into your chosen AI image generation model (e.g., Nano Banana) to produce the advertisement image based on your product and style descriptions.

5. Adjust Image Dimensions and Aspect Ratio

  • Upscale: Pass the generated image through an Upscaler node to increase its resolution (e.g., to 2048 x 2048 pixels) for high-quality output.

  • Adjust for Video: To prepare for vertical video platforms (like social media Reels), the image must be converted to the correct aspect ratio (e.g., 9x16 portrait).

    • If the image lacks sufficient background, use an Outpaint node to seamlessly extend the environment above and below the product.

    • Utilize a Crop node to finalize the 9x16 ratio and position the product effectively within the frame.

Phase 2: Video Animation and Final Compositing

This phase converts the still image into a video and adds professional branding without AI interference.

6. Generate the Video Animation

  • Input the finalized, high-resolution image into an AI video generation model (e.g., Seed Dance or Kling).

  • Use a simple, controlled prompt to guide the animation, such as: "product animation. Camera slowly moving sideways and zooming in. Product is still in focus."

7. Add Text and Logo using Compositor

  • The Compositor node is used to layer external images and text onto the AI-generated video, preserving their quality.

    • Plug the generated video as the background layer.

    • Add a separate layer for your logo image (sticker).

    • Add one or more text nodes for headlines or call-to-actions.

  • Use the Compositor's editing interface to:

    • Scale and position the logo.

    • Select the font, color, and size for all text layers.

    • Ensure all elements are positioned correctly within the video frame.

8. Export the Final Video

  • Connect the output of the Compositor node to an Export node.

  • Initiate the download process to render and save your completed, high-quality, AI-generated video advertisement.

AI Video Ads, Amazon Product Videos, Amazon Product Images, Product Videos, Product Images, Weavy, Nano BananaFrancesca Tabor