How to Make AI Testimonial Videos Without Faking Reviews

Short answer: To make an AI testimonial video, start from a real review from a real customer, get their permission, keep their words, then let an AI presenter and voice deliver the script and label the video as AI-made. AI changes how the testimonial is produced. It should not change who said it.

The test: could you show the original review, the customer's permission and the date it was given if someone asked? If yes, you are making an AI testimonial video. If not, you are making a fake review.

Why "real customer, AI delivery" is the line

AI testimonial videos save the part that is slow and expensive: casting, filming and editing. They do not replace the part that makes a testimonial worth anything: a real person who used the product and said something about it.

In the United States, the Federal Trade Commission's rule on consumer reviews and testimonials covers reviews and testimonials that misrepresent who wrote them, and the FTC names AI-generated fake reviews as an example. The FTC's own questions and answers also explain that an AI presenter delivering a testimonial is judged on whether the testimonial behind it is real, and that the way an avatar is used can still be deceptive. Other countries have their own consumer rules, so check the ones that apply to your market. This article is a production guide, not legal advice.

Step 1: Collect real reviews you are allowed to use

Good sources are product reviews on your own store, support emails, survey answers and messages from customers who offered feedback. For each one, keep:

• the original text and where it came from;
• the customer's permission to turn it into a video ad;
• the date, and whether the customer received anything for the review.

If a customer was paid, given a free product or offered a discount for their review, that connection needs to be disclosed in the ad as well.

Step 2: Turn the review into a short script without changing the meaning

Testimonial ads are short, so you will shorten the review. AI Actors and Talking Photo AI take up to 10 seconds of audio per run on VEED Fabric 1.0 and up to 9 seconds on OmniHuman 1.5, so keep each part to one or two sentences and join the parts, or pick the single strongest quote. Shorten, do not rewrite:

• keep the customer's own words and the claim they actually made;
• cut filler, keep the problem, the product and the result they described;
• do not add numbers, results or feelings the customer did not mention.

A simple shape that works for most ads: the problem they had, what they tried, what changed after using the product. For more ad script shapes, see UGC ad script templates.

Step 3: Pick how the testimonial is delivered

Choose the delivery that fits what you have:

Only use a real customer's face or voice when that customer has agreed to it. Wandura AI's voice clone works from a short audio sample, which is exactly why it needs the speaker's consent. A ready-made AI presenter reading a real customer's quote, clearly labelled, is the simplest option when you do not have that consent.

Step 4: Generate, then check the words against the review

Generate the video, then watch it with the original review open next to it. Check that:

• every sentence the presenter says is in the review, or is a faithful shortening of it;
• the presenter does not claim to be the customer unless the customer is the one on screen;
• names, product details and any results match the source.

The price of each generation is shown before you create, so you can test two or three presenters or voices for the same review and keep the one that reads most naturally.

Step 5: Label it clearly

Put the label where viewers see it, not only in the post caption. Examples you can adapt:

• "Customer review from [first name], [month year]. Voiced by an AI presenter."
• "Dramatization with an AI presenter. Words from a verified customer review."
• "AI-generated video. Review used with the customer's permission."

Many ad platforms also have their own labels or settings for AI-generated or altered content. Use them when they apply.

Step 6: Export for each placement and test

Export in the aspect ratio of each placement: 9:16 for short-video feeds, 1:1 and 16:9 for other placements. Keep the first seconds about the customer's problem, put the product early and test several real reviews against each other. For placement sizes, see TikTok ad specs.

Checklist before you publish

• The review is real, and you kept the source and the date.
• The customer agreed to the video use.
• Any payment, free product or discount for the review is disclosed.
• The script keeps the customer's words and claims.
• A face or voice of a real person is used only with their consent.
• The video says on screen that it is AI-made.