Skip to main content
  • Pricing
  1. Home
  2. YouTube summaries
  3. AI UGC ads
  4. Guide 1 of 8

How to make AI UGC adsGuide 1 of 8

AI UGC ads tutorial: how to create them start to finish

Learn how to generate authentic, high-converting user-generated content ads using artificial intelligence tools without hiring creators or shooting video.

100x Engineers

“How To Create AI UGC Ads (Full Tutorial)”9:55

Visit channel

In 30 seconds

  • UGC ads succeed because their unpolished, everyday appearance blends naturally into social feeds.
  • A structured JSON-style prompting method yields more precise visual results when designing AI influencers.
  • Combining a product photograph with an AI character image in ChatGPT creates consistent promotional stills.
  • Pairing specific spoken dialogue with concrete physical actions in AI video generators like Seedance 2.0 generates realistic video ads.

0:00Intro

The Rise of AI in User-Generated Content

0:00 to 1:09 in the video, about 1 minute

User-generated content, or UGC, refers to informal, natural-looking videos created by regular consumers sharing their experiences with products. Unlike polished commercial productions, UGC feels organic and low-effort, allowing it to stand out on social media feeds and drive higher conversion rates.

While brands traditionally pay creators hundreds of dollars per video clip, modern artificial intelligence tools can now replicate the look, feel, and voice of real people. By generating synthetic influencers and scripting their performances, marketers can drastically cut the time and expense required to produce video campaigns.

The Three-Step Production Framework

1:10 to 1:20 in the video, under a minute

Building a convincing UGC advertisement entirely with AI involves three core phases. Each phase relies on specific generative tools and prompt structures to keep character features and brand assets consistent throughout the process.

  • Step 1: Generate a custom AI influencer with a defined appearance and environment.
  • Step 2: Integrate your brand's physical product into images of the virtual influencer.
  • Step 3: Animate the influencer using a video model with synchronized speech and actions.

Step 1: Generating the Virtual Influencer

1:21 to 4:26 in the video, about 3 minutes

The first step is establishing the look of your digital spokesperson. Using image models such as ChatGPT Images 2.0 or Nano Banana Pro, you can generate realistic portraits of fictional people. The key to consistency is structured prompting, often referred to as key-value or JSON prompting.

By defining explicit parameters rather than writing loose sentences, you give the generator clear constraints that produce more authentic, less artificial portraits.

  • Gender and Age: Tailor these attributes to the product's demographic, such as choosing younger adults for energy drinks or older adults for wellness items.
  • Ethnicity and Setting: Define the subject's background and surroundings, like a gym or bedroom.
  • Camera Style and Placement: Specify handheld angles, phone camera aesthetics, or stationary tabletop setups.
  • Framing and Ratio: Set vertical aspect ratios (typically 9:16 for TikTok and Instagram) and shot coverage such as medium close-up or full-body.

Step 2: Integrating the Product with the Influencer

4:27 to 5:23 in the video, under a minute

Once you have an influencer image you like, the next phase is showing them interacting with your item. You need a high-resolution, well-lit photo of the product where packaging details and labels remain clear.

By uploading both the influencer image and the product photo to an image-capable AI tool, you can instruct the model to merge them. Requesting variations—such as holding the bottle, placing it on a table, or applying the contents—gives you a library of starting points for the final video.

Step 3: Generating Dialogue, Movement, and Video

5:24 to 9:37 in the video, about 4 minutes

The final step turns the still composite image into an animated, speaking advertisement using advanced AI video generators like Seedance 2.0. The video model requires clear instructions about how the persona talks and behaves on screen.

A dedicated video prompting framework pairs vocal style with deliberate physical cues. Specifying parameters such as accent, emotional vibe, camera motion, and eye direction ensures the output does not feel stiff or detached.

Crucially, the script should be structured as paired lists: each line of spoken dialogue must be matched with a corresponding action, such as lifting the product, examining the bottle, or showing the label directly to the lens. Creators can also generate supplementary B-roll shots and edit them together into a complete social ad.

Final Thoughts

AI video generation has reached a point where authentic, handheld UGC commercials can be designed from scratch in minutes. By using structured prompts for both character creation and video directing, marketers can quickly test multiple concepts and spokespeople without the delays of physical filming.

Do this yourself

Make your first UGC-style ad

The video's three steps are a creator image, the product in their hands, then the final clip. In the UGC Ad Creator you pick an avatar, add your script or audio, and generate a lip-synced ad clip you can test.

Questions

Which AI tools are used to create AI UGC ads?

Image generation and image blending can be handled with tools like ChatGPT Images 2.0 or Nano Banana Pro, while video generation and voice animation are performed using models like Seedance 2.0 on platforms such as Higgsfield or OpenArt.

What is JSON prompting for image generation?

JSON prompting is a structured prompting style that organizes instructions into clear key-value pairs (such as specifying age, style, camera placement, and aspect ratio) to produce more accurate and repeatable AI visuals.

How do you ensure the AI influencer interacts naturally with a product?

You supply the video generator with a detailed script that links each spoken sentence directly to a physical motion, such as raising the bottle or displaying the label to the camera.

This is a summary of a public video by 100x Engineers. The method and the demo are theirs: watch the original. Oakgen wrote the summary.