AI UGC for Facebook Ads: Prospecting, Retargeting and Meta's AI Label
AI UGC for Facebook ads means generating photos and short videos of your real product in believable scenes, often held or used by a consistent AI person, then running them in feed, Stories, Reels and carousel placements. It is most useful where Meta advertisers usually run short of creative: different concepts for prospecting and fresh angles for retargeting. The caveat to plan for: Meta can show an “AI info” label on ads it detects were made or heavily edited with generative AI, including third-party tools.
3 min read
Most Meta accounts are not short on ad ideas. They are short on finished, distinct images to test those ideas with, and the gap shows up twice: when you need new concepts for cold audiences, and when your retargeting audience has seen the same three product shots for a month. This guide covers what to generate for each, how Meta labels AI-made ads, and the uses Meta's own policies and the FTC already rule out.
How Meta's “AI info” label works on ads
Meta's Business Help Center page on how AI-generated images in ads are identified and labeled describes the current behavior:
- An “AI info” label can appear in the ad's three-dot menu under “About this ad”, and sometimes next to the “Sponsored” label.
- It is applied to ad images created or significantly edited with Meta's own generative AI creative tools, and Meta said in February 2025 that it would also detect ads made with third-party AI tools through industry-standard signals.
- Minor edits such as resizing or color correction do not trigger it.
- Ads about social issues, elections or politics have a separate, stricter rule: advertisers must disclose realistic AI-made or AI-altered media themselves.
For a product ad, the practical takeaway is simple: assume the label may appear and make creative that still works if it does. An AI scene of your actual product on a kitchen counter loses nothing when a viewer learns the scene was generated. A “customer” who turns out not to exist loses everything.
The frames to generate for each placement
| Placement | Generate at | What tends to work as AI UGC |
|---|---|---|
| Facebook and Instagram feed | 4:5 portrait (1080x1350) | Person holding or using the product, product clearly visible in the top two thirds |
| Stories and Reels | 9:16 vertical (1080x1920) | Selfie-angle or mirror shots; a short talking clip from the same person |
| Carousel | 1:1 square, 2 to 10 cards | One person across cards: unboxing, in use, detail, result you can actually show |
| Advantage+ catalog / collection hero | 1:1 or 4:5 | Lifestyle hero of a best seller above the catalog grid |
Generate each ratio as its own composition rather than cropping one landscape image three ways; crops tend to cut off the product or the face. Leave the bottom of 9:16 frames uncluttered, since Meta places the call-to-action and caption there.
Prospecting: give Meta different ideas, not ten crops
With broad targeting and Advantage+ placements, the creative does most of the targeting. Ten near-identical images of the same person in the same kitchen give the system one idea to test, not ten. Vary the things a buyer would notice:
- Who: two or three AI Experts that match different buyer groups, each kept consistent across their own images.
- Where: the situations the product is actually used in (commute, gym bag, bathroom shelf, desk), not generic “lifestyle”.
- Why: one concept per reason to buy. If the reasons are price, durability and how it looks, make one image that argues each.
- Format: a still, a carousel and a short video of the same concept.
For Advantage+ specifics, see the Advantage+ campaigns guide; for the platform setup in ppl.studio, see ppl.studio for Meta Ads.
Retargeting: answer the reason they did not buy
A retargeting audience has already seen your product. Showing them the same prospecting image again adds nothing. Instead, guess the objection that stopped them and make creative that answers it:
| Likely objection | Creative that answers it | What not to fake |
|---|---|---|
| Not sure it is worth the price | The product in a setting that shows quality and scale: close detail, real materials, next to everyday objects for size | Fake price comparisons or 'was' prices |
| Not sure about fit, size or quality | Several angles of the real product, worn or held by people of different builds, honest about size | Fit or results the product does not deliver |
| Not sure about the brand | Real reviews and real customer photos, with permission; AI scenes only for product context | AI faces presented as reviewers |
| Just got distracted | A fresh scene of the exact product they viewed, a new season or setting | Urgency that is not real ('only 2 left') |
Sequencing helps here. Because an AI Expert keeps one face, the same person can appear in a short run: the product in use, then a close detail, then a seasonal scene. Treat it as a recurring character in your ads, not as a customer.
What Meta and the FTC already rule out
- Invented testimonials. The FTC's consumer reviews and testimonials rule (16 CFR Part 465, effective October 2024) prohibits testimonials that falsely claim to come from a real person, explicitly including AI-generated ones. An AI person can demonstrate your product; it cannot be your reviewer.
- Before-and-after images for health and appearance. Meta's advertising standards already restrict these for weight loss and similar products, and generated “after” images are fake evidence anywhere.
- Partnership ads with a creator who does not exist. Meta's partnership ads run through a real creator's account with their permission. An AI persona has no account to lend; run AI UGC from your own Page.
A weekly routine that fits a small team
- Monday: look at last week's ads. Note which concept (not which image) won.
- Generate three new concepts and two variations of the winner, in 4:5 and 9:16. The Product UGC Photos workflow produces a lifestyle set from one product upload.
- Turn the strongest still into a short spoken clip with Animate, or use the Creator UGC Video workflow for a scripted one.
- Refresh retargeting with one new scene per objection above.
- Pause what has clearly lost; keep what is still working. Do not rotate creative for its own sake.
ppl.studio costs $1.99 per week (pricing), and images are billed by Google on your own Gemini API key. For multi-frame creative see the carousel guide; for organic Facebook and Instagram, see AI UGC for social media.
Ad-ready scenes for one product
One AI Expert and one real product, placed in different scenes by ppl.studio. Same face, same product, every shot.






Generate Meta ad concepts from your own product photos
Upload a product, choose an AI Expert, and generate feed, Stories and carousel frames in one session. 10 free photos, no credit card.
Start free with ppl.studio10 free photos · no credit card required
Founder of ppl.studio. Building AI tools for product marketing teams who need visual content at scale without the production overhead.