An AI UGC tutorial ad works best when it begins with one approved composite: a believable creator image plus an exact product reference. I treat that still as the gate. If the label, cap, fingers, or product scale is wrong before animation, I stop there—motion will amplify the error, not reliably repair it.
This workflow avoids the biggest mistake in AI product ads: asking one prompt to invent the person, product, scene, action, and voice at the same time.
The Showcase Pattern: Person First, Product Second
For product UGC, separate the creator from the product during asset prep.
Create the person image with empty hands and no product visible. Then create a product angle grid from the real product reference. This gives the video model two clean jobs: keep the person consistent and keep the product recognizable.
Use this reference stack:
| Asset | What it should show | Why it matters |
|---|---|---|
| Creator image | Original adult creator, clear face, natural setting, empty hands | Keeps the presenter usable across many ad versions |
| Product angle grid | Front, back, side, detail, macro, lifestyle, scale, hero angle | Gives the model enough views to preserve packaging |
| Seedance prompt | Shot order, product placement, camera behavior, dialogue | Turns the two references into a vertical UGC clip |
| TTS audio | Short voice line generated with DomoAI Text to Speech | Gives Talking Avatar clean audio for lip sync |
This is also why the first image should not already show the product. If the creator is already holding a random generated object, the product can drift before the real demo starts.
Build The Creator Reference
Use DomoAI Text to Image or an uploaded creator photo. The goal is not a perfect ad still. The goal is a reusable presenter reference.
For a product demo, write the creator prompt with three stability rules:
- Empty hands.
- No product visible.
- A real setting that matches the product category.
Showcase prompt:
Original adult woman startup founder in a clean modern office nook, black turtleneck, notebook and laptop behind her, hands empty, not holding anything, no product visible, natural daylight, polished but realistic UGC portrait.
That prompt gives you a believable founder in a workspace. It does not ask the image model to invent the purifier yet. That comes next.
For a fitness product, use the same logic:
Front camera selfie of an original adult man after a morning run, standing near a sunny park path, athletic shirt, light sweat, relaxed breathing, hands empty at his sides, not holding anything, no product visible, natural iPhone UGC style.
The words hands empty, not holding anything, and no product visible look repetitive, but they help. They stop the source image from inventing a fake bottle before you upload the real one.
Prepare the Creator and Product Assets
An AI UGC tutorial ad works best when you prepare two assets: a clean creator reference and an exact product reference. Use AI Image Editing—Nano Banana Pro is the best fit for multi-reference work—to combine them into one approved source still. Review the face, hands, label, packaging, and scale before animating that still with Image to Video. Use Talking Avatar with DomoAI Text to Speech for a separate face-led narration shot when precise speech timing matters.
Create at least two approved composites: one presenter-led frame and one product-detail frame. Keep important label text away from hands and motion-heavy areas.
Showcase 1: Founder Desk Product Demo
This version is a founder-style UGC review for a portable air purifier. It uses two images: the founder reference and the product angle grid.
Use Image to Video with Seedance 2.0 when you want the clip to include product motion, creator framing, and a short multi-shot sequence.
Seedance prompt:
Input: Approved founder-with-product composite still.
8-second 9:16 phone-style UGC clip. Keep the founder identity and product packaging unchanged. Start on a desk-level medium shot, cut to a front-visible product close-up, then return to the founder. Natural daylight, subtle hand movement, stable label, no generated captions, no spoken testimonial.Generate narration separately with TTS: “For a small workspace, a compact setup matters. This fits neatly beside a laptop without taking over the desk.”
Why this works:
- It gives the model a clear shot order.
- It tells the product when to enter the scene.
- It asks for a front-visible product moment.
- It avoids medical or performance claims.
- It makes the voice sound like UGC, not a studio ad.
The line is also safe for a product demo. It says the product fits the workspace and looks clean on the desk. It does not claim the purifier removes a specific pollutant or creates a health result.
Showcase 2: Runner Voiceover And Product Clip
This version is for electrolyte drops. It can work in two ways.
Use Seedance when you want the runner to hold the bottle in the product clip. Use Talking Avatar when you want cleaner lip sync from a face close-up and separate product visuals.
Seedance prompt:
Input: Approved runner-with-product composite still.
6-second 9:16 phone-style product clip. Keep the creator identity and bottle packaging unchanged. Small nod, bottle held near shoulder height, label facing camera, no pouring, mixing, drinking, generated captions, or spoken testimonial.Generate narration separately with TTS: “Electrolytes are easy to overlook after a short run. This small bottle is easy to carry in a gym bag.”
This prompt keeps the action small. It does not ask for pouring, mixing, or drinking. That is useful when the product is small and the label matters.
For a more controlled voice track, generate the line in DomoAI Text to Speech first. Then use the TTS output as the audio for Talking Avatar.
Talking Avatar prompt:
Use @Image and @Audio. Fitness creator close-up, accurate lip sync, natural blink rhythm, no product visible.
This version separates the product clip from the spoken face clip. That often looks cleaner in ads. The product can stay readable in Seedance, while Talking Avatar handles the close-up narration from the generated TTS audio.
Seedance Or Talking Avatar?
Use the tool based on the job of the shot.
| Shot job | Better fit | Why |
|---|---|---|
| Creator holds product and turns it to camera | Seedance / Image to Video | Animate an approved composite while keeping the product and presenter consistent |
| Product close-up on desk, counter, bag, or hand | Seedance / Image to Video | The prompt can focus on label stability and camera movement |
| Face close-up with reliable speech timing | Talking Avatar + DomoAI TTS | The audio drives the lip sync more directly |
| No-face voiceover with product visuals | DomoAI TTS + product clips | The voice explains while the product remains the visual focus |
| Green-screen presenter for later editing | Talking Avatar with Screen Keying | You can composite the presenter outside DomoAI |
Do not force every ad into one format. A good UGC tutorial ad may use a Seedance product shot, a Talking Avatar line, captions, and a final product close-up in the same edit.
Turn The Showcase Into A Tutorial Ad
After the first clip works, cut it into a small tutorial structure.
For the air purifier:
Hook: My desk is tiny, so every product has to earn its space.
Demo: Show the purifier on the desk beside a laptop.
Proof: Show the product scale, front view, and clean desk placement.
CTA: Save this if you're building a quieter workspace setup.
For the electrolyte drops:
Hook: I used to skip electrolytes after short runs.
Demo: Show the small bottle near the shoulder with the label visible.
Proof: Show the bottle beside a gym bag or water bottle.
CTA: Keep this in your bag before your next run.
These are tutorial ads, not hard-sell testimonials. They teach a small behavior: how the product fits into a workspace or a running routine.
Keep Captions Outside The Generated Clip
Most viewers see UGC ads muted first. Add captions in your editor after generation so the text stays readable.
Use captions as the tutorial spine:
- Tiny desk setup.
- Product stays small.
- Front label visible.
- Quick desk close-up.
- Save for your workspace reset.
Or for the runner:
- Post-run routine.
- Small bottle.
- No mixing shot needed.
- Easy to carry.
- Save for your gym bag.
Avoid putting these captions inside the generation prompt. Generated text can bend, misspell, or cover the product. Add caption text after the video is approved.
Disclose and Substantiate the Ad
Treat AI UGC as advertising, not as an independent customer testimonial. If the avatar did not use the product, do not script first-person experience claims or imply that it is a real customer. Make the commercial nature of the content clear, disclose material brand connections where required, and substantiate express or implied performance claims before launch. Platform AI-label requirements vary, so check the destination platform’s current policy.
Use a Stop/Go Gate Before Animation
I do not animate a merely “close enough” product still. I score the composite first, because the costliest failures are usually visible before the video stage.
| Check | Go | Stop and fix |
|---|---|---|
| Product identity | Label, color, cap, and silhouette match | Any brand-defining detail changes |
| Scale | Product size looks plausible in the hand or scene | The object reads as miniature or oversized |
| Contact | Fingers and product edges meet cleanly | Hands melt into the package |
| Text safety | Important text stays unobstructed | A finger or crop hides the proof |
| Caption room | One clear area remains for post-production text | The subject fills every usable zone |
Then I rate the action risk. Holding, tilting, and placing are low risk. Opening and handoffs are medium risk. Pouring, spraying, or mixing are high risk because the product, hand, and material all need to stay coherent. The product-integrity guide explains the same problem in more depth, and the packaging-consistency workflow gives it a concrete ad context.
For testing, I change one variable at a time: opening frame for two-second hold, proof order for completion, or CTA for click intent. I keep the approved composite fixed. When the final crop is ready, I use Video Upscaler only after the product details have passed review.
Build Three Test Variants
Keep the same creator and product references. Change only one creative variable per test.
| Variant | Change | Keep stable |
|---|---|---|
| Product-first | Open on the product close-up | Creator, product grid, dialogue |
| Creator-first | Open on the person speaking | Product reference, caption style, CTA |
| Use-case-first | Open on the setting or problem | Product shot, voice line, final frame |
For the founder clip, test whether the desk close-up or the founder face works better as the first frame. For the runner clip, test whether the product near the shoulder or the sunny park selfie earns the faster stop.
Do not change the actor, product angle, voice, captions, and CTA at once. You will not know what improved the ad.
Review Fixes Before Export
Review the first generation like a muted mobile viewer. The product, proof, and next step should still read before the voice matters.
The product looks wrong
Regenerate or edit the product angle grid. Add exact packaging, label, cap, color, and scale rules before making more video versions.
The creator feels too polished
Use phone-style language in the prompt. Ask for natural daylight, relaxed posture, and a phone-recorded voice rather than studio lighting or polished ad VO.
The lip sync is not clean enough
Use DomoAI Text to Speech to create a short audio line, then feed that audio into Talking Avatar. Keep the avatar as a close-up and move product handling into separate Seedance shots.
The product action gets messy
Remove complex actions. Use holding, tilting, placing, pointing, or a desk close-up. Avoid pouring, mixing, spraying, or opening unless the product action is essential.
The ad feels too salesy
Move the CTA later. Let the first half show the use case. A tutorial ad should feel like a creator explaining a small routine.
FAQ
What is an AI UGC tutorial ad?
An AI UGC tutorial ad is a short creator-style product video that teaches one use case before asking for action.
Can I make one from only a product photo?
Yes. Create a product angle grid from the product photo, generate or upload a creator reference, then make the product demo clip.
Can Talking Avatar use DomoAI Text to Speech audio?
Yes. Generate the voice with DomoAI Text to Speech, then use that audio in Talking Avatar for a lip-synced creator close-up.
Should the product appear in the creator reference image?
Prepare and approve the creator and product together in one composite still before animation. Check hands, packaging, label placement, and scale in the still rather than asking the motion model to assemble them.
Should every UGC ad use a talking avatar?
No. Use Talking Avatar when speech timing matters. Use Seedance product clips when the product, handoff, desk shot, or lifestyle context matters more.
How long should the first version be?
Start with 6-8 seconds for a showcase clip, then expand into a 15-25 second ad if the product and creator look stable.
Create the creator reference, build the product grid, generate one Seedance product clip, and add TTS-driven Talking Avatar only where the voice needs precise lip sync.
For paid campaign planning, use AI Ads Solution to connect each tutorial clip to a hook, offer, and CTA. For a no-face variant, Product Explainer Video Without Showing Your Face can turn the same product references into a voiceover-led explainer.
