UGC-Style Product Testimonial Video Prompt
This prompt generates the UGC testimonial format dominating TikTok ads: a real-feeling customer holding your product, talking enthusiastically in her kitchen, with the slight shake and imperfect framing that signals 'authentic' to viewers' brains. Describe the person, product, and setting — the deliberate lack of polish is the entire strategy.
The TikTok-native ad format: shaky, authentic, high-converting — a customer testimonial that doesn't look like an ad.
UGC-style customer testimonial video, 10 seconds, shot on a phone: [PERSON — e.g. a woman in her late 20s, no makeup, casual hoodie] holding [PRODUCT] up to the camera in [SETTING — e.g. her kitchen, morning light], speaking enthusiastically and gesturing naturally. Slight handheld shake, authentic imperfect framing (a little too close), genuine smile, mid-sentence energy like she just had to tell someone. Vertical 9:16, natural lighting, no studio polish, no text overlays, photorealistic.
See what this prompt creates
Here's the paradox running paid social in 2026: the worse the production value, the better the conversion — up to a point. UGC-style creative outperforms polished brand spots because viewers' brains file it under "person" instead of "ad," and the scroll doesn't stop for ads. This prompt reverse-engineers that: the handheld shake, the too-close framing, the mid-sentence energy are all deliberate craft masquerading as accident.
The critical insight is that authenticity is a set of visual signals, and signals can be directed. "Slight handheld shake" (not seasickness), "a little too close" (not a nostril cam), "morning light" (not a ring light) — each phrase walks the line between real and unwatchable. Push any of them too far and the illusion breaks the other way: viewers smell the try-hard. The prompt is calibrated to the middle of that line.
Getting the best result
- Cast the right "customer." No makeup, casual clothes, real-person energy — the testimonial fails the moment the presenter looks like a model. Relatability is the product here.
- Vertical, always. 9:16, full stop. This format lives on TikTok, Reels, and Shorts; generating horizontal and cropping wastes half your frame.
- Direct the energy, not the words. "Mid-sentence energy like she just had to tell someone" — that urgency is what stops thumbs. A calm testimonial is a skipped testimonial.
- Pair with real copy. Generate the visual here, add your hook text and captions in CapCut. AI visuals + human copy beats AI everything.
3 variations to try
Unboxing: "just opened the package on camera, tissue paper visible, genuine first reaction" — the highest-retention UGC sub-format.
Before/after: "split screen energy: tired morning face left, glowing after using [PRODUCT] right" for beauty and wellness products.
Car testimonial: "filmed in a parked car, seatbelt on, quick enthusiastic review" — the oddly high-converting classic of the genre.
How to use this prompt
- Copy it — hit the copy button above to grab the raw prompt text.
- Fill the brackets — replace anything in [BRACKETS] with your own details.
- Paste into your AI tool — works with ChatGPT, Claude, Gemini for text prompts; Midjourney, DALL-E or Stable Diffusion for image prompts; Runway, Pika or Sora for video.
- Iterate — tweak one bracket at a time until the output is exactly right.
Frequently asked questions
Isn't AI UGC deceptive advertising?
It can be — which is why disclosure matters. Many ad platforms now require AI-generated content labels, and several jurisdictions treat undisclosed synthetic testimonials as deceptive. Use the format for its visual language, label it per platform rules, and never fabricate claims about real customers.
Why does polished video underperform UGC?
Pattern interruption. Viewers have banner-blindness for anything that looks like an ad; UGC's visual signals (shake, casual framing, real rooms) file it under 'person sharing something' instead. The scroll stops, the message lands. It's psychology, not production value.
How long should a UGC ad be?
Ten seconds is the sweet spot for the generated clip — hook, product, reaction. Cut multiple 10s generations into 25–30s spots in editing. Single generations beyond ~10s tend to drift in energy and coherence.
Can I use the same avatar across multiple ads?
Consistency across generations is still hard — the 'same' person will drift. For a recurring brand character, generate one strong reference and use image-to-video with character reference features where your tool supports them, keeping each spot short.
