Podcast Style UGC

This tutorial walks you through creating a complete video ad featuring a fictional influencer promoting a product. The ad uses a before-and-after transformation format, which works well for health, beauty, and lifestyle products. You will generate all the images, video, and voice from scratch using AI, then edit everything together into a polished short-form ad.

Step 1: Generate the Base Influencer Image

Create the character who will appear throughout your ad by generating a base image using the influencer mode.

• Open the image generator and select Influencer Image mode.
• Describe your character in plain words. For example: "older woman."
• Wait for the generation to complete.
• Review the result. You will refine it in the next step.

Step 2: Refine the Influencer Image with NanoBanana

Make the base image feel more natural and relaxed by editing it with NanoBanana.

• Open NanoBanana Pro and select your generated image.
• Describe the changes you want. For example: "Make her more relaxed and casual, wearing a crop top, sitting on a laid-back podcast set."
• Generate the updated image and confirm it looks natural and usable.

Step 3: Clone a Reusable Voice

Set up a consistent voice for your ad by cloning a voice from a previous Sora video generation.

• Find a Sora generation with a voice you like.
• Clone the voice using the Voice Cloning feature so you can reuse it across this ad and future projects.

Step 4: Generate the Main Video with Sora

Create the talking-head video that will be the core of your ad.

• Open the Sora 2 Pro video generator. Select sixteen seconds and pro mode.
• Describe the image (what the scene looks like).
• Describe what should be spoken in the video.
• Use the Prompt Optimizer tool to automatically generate the Sora prompt from your plain-language description.
• Start the generation. Expect about twenty minutes of wait time for a sixteen-second clip.

Step 5: Create a "Before" Selfie Image

Generate a selfie that shows the character before their transformation.

• Describe the selfie scene using NanoBanana Pro: the woman in a bathroom, wearing pajamas, taking a mirror selfie, looking natural.
• If needed, push the model to exaggerate problem areas (for example, more visible bald spots) to create a stronger contrast for the transformation.
• Generate the image.

Step 6: Generate the "After" Transformation Photo

Create the final aspirational image showing the character with the product.

• Describe the scene: the woman sitting in a luxury car (for example, a Bentley) holding the product.
• Turn on the Google Search Grounding toggle when referencing real-world items like car brands. This helps the AI find accurate reference material like car interiors.
• Generate the image and confirm it looks polished.

Step 7: Find Background Music

Select a music track that fits the mood of your ad.

• Use the Stock Music Search tool to browse royalty-free tracks.
• Preview tracks and select one that matches your ad's tone.
• Add it to your project.

Step 8: Edit and Polish the Final Ad

Bring all your assets together in the AI Video Editor and handle any AI artifacts.

• Click Export to Editor and upload all generated assets (video, images, music).
• Watch the Sora video carefully for AI artifacts like extra fingers or unnatural hand movements.
• Place B-roll images (the selfie, the product photo) strategically on top of any problem moments in the video.
• Use image cutaways to cover sections where hands morph or look unnatural.
• Add and polish subtitles.
• Zoom in on key moments for emphasis.
• Rewatch the full video multiple times to catch anything you missed.
• Render the final ad.

Final Result

The finished ad is a short-form UGC-style transformation video. It features a fictional older woman influencer sharing a before-and-after story with a product. The ad includes a talking-head video segment, a "before" selfie, and an aspirational "after" photo with the product in a luxury setting. Background music and subtitles tie everything together into a polished, scroll-stopping ad ready for social media.

Models Used

• Influencer Image — Generated the base character image for the fictional influencer.
• NanoBanana Pro — Used to refine the influencer image, generate the "before" selfie, and create the "after" transformation photo.
• Sora 2 Pro — Generated the sixteen-second talking-head video clip that forms the core of the ad.
• Prompt Optimizer — Converted plain-language descriptions into optimized Sora prompts.