How to Create Scroll-Stopping AI UGC Videos in 3 Easy Steps

User-generated content (UGC) is everywhere right now. In a sea of high-production, heavily edited commercials, a natural-looking, low-effort video stands out because it feels authentic and relatable.

Traditionally, brands pay creators anywhere from $100 to $500 for a single UGC video. But the landscape is shifting. AI can now generate highly convincing UGC ads—featuring avatars that look like real people and voiceovers that don't sound robotic—in the time it takes your coffee to get cold.

Whether you are a marketer, brand owner, or creator looking to scale, here is the complete three-step workflow to create high-converting UGC ads at a fraction of the cost.

--------------------------------------------------------------------------------

Step 1: Design Your AI Influencer 🎨

Your first goal is to nail down the look of your AI creator. To do this, you can use powerful image generation models like GPT Image 2 or Nano Banana Pro, which are accessible through platforms like Open Art and Hicksfield.

To get the most realistic outputs, you should use a highly structured technique called "JSON prompting." This involves providing specific keys and values to the AI. Start your prompt with "Create a UGC influencer" and include the following parameters:

• Age & Gender: Tailor this specifically to your product. If you are selling an energy drink for Gen Z, generate a 20-25 year old; if you are promoting a hearing aid, opt for a 50-60 year old.

• Ethnicity & Additional Details: Define their background and add granular details like curly hair or specific eye colors.

• Style & Coverage: Keep the style "natural" or "realistic" (rather than stylistic), and define if you want a facial close-up, a half-body shot, or a full-body shot.

• Camera Placement & Aspect Ratio: Specify if the influencer is holding a selfie stick or if the camera is stationary. Since UGC thrives on TikTok and Instagram, use a 9:16 aspect ratio.

Pro Tip: If you aren't happy with your first generation, keep iterating and prompting until you find the perfect face for your brand.

Step 2: Place Your Product 📸

Once you have your influencer, it is time to put your product in their hands.

Start with a high-resolution, well-lit picture of your product to ensure that any text or details don't get lost in the generation process. Next, head over to ChatGPT and upload both your AI influencer image and your product photo.

Use a simple prompt: "Combine these two images, I want this person holding this product".

It is highly recommended to generate a few different scenarios—like the influencer holding the product or the product sitting on a table in front of them—so you have multiple starting points for your video.

Step 3: Bring the Ad to Life 🎬

Now for the magic. To turn your static image into a talking, moving ad, you will use Seed Dance 2.0, which is currently the leading video model for this task (accessible via Hicksfield or Open Art).

Upload your combined product-and-influencer image and use this second prompting framework to control the video output:

• Accent & Vibe/Mood: Tell the AI exactly how the person should sound (e.g., French, British, Indian) and what their personality should be (e.g., enthusiastic, casual, or semi-formal).

• Camera Motion & Eye Direction: Do you want them walking with a selfie camera or standing still? Should they be making direct eye contact with the lens or looking down the street?

• Dialogue / Action List: This is the most crucial step for realism. You must provide a structured list that pairs specific dialogue with specific actions.

Here is an example of how to format your dialogue and action list:

1. Dialogue: "I absolutely love this hair serum." / Action: The lady takes the hair serum from her table and has it in her hand.

2. Dialogue: "I've been using it for 3 months and it's better than anything I've ever used." / Action: She is looking at the product and saying these lines.

3. Dialogue: "I absolutely recommend this to everyone." / Action: She is showing the product to the camera.

By strictly pairing the dialogue with an action, the AI nails the timing perfectly. If you want to get even more creative, you can generate additional B-roll shots while your avatar is talking and stitch them together in editing