ppl.studio

AI for Customer Testimonial Visuals: Illustrate Real Reviews, Never Invent Them

You can use AI to make testimonial visuals for real reviews: the customer's actual words, used with their permission, set on a product scene, a quote card or a clearly labeled video. You cannot use AI to make the customer or the review. Since October 21, 2024, the FTC's rule on consumer reviews and testimonials (16 CFR Part 465) prohibits reviews and testimonials that misrepresent they come from someone who doesn't exist, from someone with no real experience of the product, or that misrepresent that experience—and the FTC names AI-generated fakes as covered. Everything below works inside that line.

5 min read

Quote card with a real customer review set beside an AI-generated product scene, labeled as an illustration

Most brands have the opposite problem from the one AI tools are usually sold for. They have reviews—dozens or thousands of real ones sitting in Judge.me, Yotpo, Amazon or Google—and no good way to show them. A five-star quote on a white rectangle doesn't stop a scroll. AI is useful here as the art department: it makes the scene, the background and the format. The words and the person stay real.


What the 2024 rule bans

The FTC's August 2024 announcement of the final rule lists the practices it covers:

  • Fake or false reviews and testimonials—ones that misrepresent they are by someone who does not exist (the FTC calls out AI-generated fakes), by someone who had no actual experience with the business or product, or that misrepresent the reviewer's experience. Businesses may not write, create or sell them, and may not buy or spread them when they knew or should have known they were fake.
  • Buying reviews conditioned on sentiment—paying or giving incentives for positive (or negative) reviews.
  • Undisclosed insider reviews and testimonials from owners, staff or their families.
  • Company-controlled “independent” review sites.
  • Review suppression, including implying the displayed reviews are representative when negative ones were hidden.
  • Fake social indicators such as bought followers or views.

The FTC can seek civil penalties for knowing violations, and in December 2025 it sent warning letters to ten companies about possible violations. Its Q&A on the rule adds two points that matter for AI: there is no blanket ban on AI-generated avatars in marketing, and an avatar used as a testimonialist is covered when the underlying testimonial is fake or false. It also notes that using actors to portray testimonialists can be deceptive under the FTC Act itself, which is why labeling matters even when the words are real.


Where the line sits in practice

FineNot fine
A real review, quoted accurately, on an AI scene of the product alone (a mug on a desk, a serum on a shelf)A review written by a person, a tool or an AI model on behalf of a customer who didn't write it
A real customer's own photo, with permission, cleaned up for lighting or croppedAn AI face placed beside a real quote so it reads as the reviewer
A real review read on camera by an AI presenter, labeled on screen as a customer review read by an AI presenterA “composite customer” who sums up many reviews in first person
A product-in-hand scene showing real size, next to a review that mentions sizeAI before-and-after results attributed to a customer, or to anyone
A summary of themes from real reviews (“customers mention the lid most”) when it is accurateTrimming a mixed review into a rave, or showing only five-star quotes while implying they are typical

Permission and paperwork

Posting a review on your store is not the same as agreeing to appear in an ad. Before a testimonial goes into paid creative or onto a hero banner:

  • Get written consent for the specific use: quote, first name or initials, photo if any, and the channels (site, email, paid social).
  • Keep the trail. Store each asset with the review's source, date and ID. If a customer later asks to be removed, you can find every image their words are on.
  • Disclose connections. If the reviewer got the product free, a discount or anything else for reviewing, the FTC's Endorsement Guides expect that to be disclosed, and any incentive can't depend on the review being positive.
  • Quote, don't improve. Fixing a typo is fine. Changing what the customer meant is not.

Formats that work with real testimonials

Quote card on a product scene

The workhorse. Generate the product in the setting the customer describes (“I keep it by the sink” becomes a bathroom-counter scene) with clear space for the quote, and set the text yourself rather than asking the image model to write it—models misspell. Build a small library of scenes per product in your brand's palette so cards look like a series. Upload the real product to the Props Library so the scene shows your item, not a lookalike.

Scale and use shots next to the review

Many reviews answer the question “how big is it?” or “does it fit?” A product-in-hand or on-a-shelf image placed next to that review makes the answer visible. Keep the hand anonymous (no face) and the proportions true.

Video testimonials

Real customer video is best. When you only have text, the testimonial video workflow can put an AI presenter on screen. Use it either to read a real, permitted review with an on-screen label saying so, or as an openly scripted brand ad where the presenter speaks for the brand. Don't script it as an invented customer's first-person story. The Animate guide covers the talking-head setup.


Review galleries and “customer photos”

If your product page has a section called customer photos or reviews with photos, everything in it must come from customers. Put AI lifestyle images in the main brand gallery, where shoppers expect brand photography, and never mix them into a review widget. The same goes for marketplace listings: brand images go in the brand slots, and the review section belongs to buyers. For gallery planning see Shopify product page images.


New product, no reviews yet

This is the moment the temptation is strongest, and the honest options are well known:

  • Sample before launch. Send product to early customers or a waitlist, ask for honest reviews, disclose that they got it free, and don't make the gift depend on a good rating.
  • Lean on what is true now. The founder's reason for making it, ingredient or material specs, warranty and returns, press if you have it.
  • Show the product, not a fake audience. AI lifestyle images of the product in use are brand photography; they don't claim anyone bought it. The product UGC photos workflow makes a six-shot set.
  • Import real reviews you already have from another channel (Amazon, wholesale accounts) where your review app allows it, labeled with their source.

For AI-content labeling more broadly, see AI-generated content disclosure and FTC guidelines. If you work with affiliates or ambassadors, the same rules from their side are in AI UGC for affiliate marketing. Plans are on the pricing page.

A plain-language summary, not legal advice. Sources: FTC, Final Rule Banning Fake Reviews and Testimonials (August 2024); FTC, Consumer Reviews and Testimonials Rule: Questions and Answers; FTC, Endorsement Guides: What People Are Asking.

A testimonial-style video

Testimonial video: a sunrise wake-up light after two weeksA review-style testimonial of a wake-up light, made with the Testimonial Video workflow.

Put your real reviews on better visuals

Upload your product, generate the scenes and quote-card backgrounds, and pair them with testimonials your customers gave you. $1.99 per week, images billed by Google on your own Gemini key.

Start free with ppl.studio

10 free photos · no credit card required

M

Max Zeshut

Founder of ppl.studio. Building AI tools for product marketing teams who need visual content at scale without the production overhead.