How Advinky makes a UGC video ad, step by step

A UGC-style ad uses a presenter talking to camera about a product. Advinky uses a synthetic presenter and an approved script to create this format. Here is how a creative direction becomes a clip you can review and edit.

Step 1 — Direction in, script out

You start with a direction. It can be a full idea, a one-liner, or one of the AI-detected trends Advinky surfaces for your category. You also choose the length (five to fifteen seconds), the language, and optionally a creator style.

The scriptwriter takes that plus your Brand DNA and writes one talking-head script. It is told to behave like a person with a phone, not a film director: a hook that lands in the first two or three seconds, a reason to care, the product, and a spoken call to action that says the brand name aloud. The scriptwriter is instructed to preserve your brand name. Review both the written script and the generated pronunciation before publishing.

Two things the scriptwriter is not allowed to do: invent claims the Brand DNA does not support, and write more words than the clip can hold. That second rule turned out to matter enormously, and it has its own article.

Step 2 — Casting the creator

A UGC ad is only as believable as the person in it. You have two options:

• Pick from the library. Filter AI creators by age range, look and vibe. Each provides a reusable visual reference to guide the generated presenter.
• Let the AI cast. Leave the choice open, or describe the person you want, and the AI casts a creator that fits your audience and brand.

The selected creator’s reference guides their face, hair, apparent age and wardrobe. AI output can still vary, so check identity consistency in the finished clip.

Step 3 — You approve the words

Nothing is rendered until you say so. The script is shown in full, beat by beat, with the spoken lines separate from the visual directions. You can revise it in conversation — "make the hook a question", "mention the free trial" — then review the revised script before approving it.

Script approval gives you a chance to correct the message before paying for video generation. The approved dialogue guides the model, but it does not guarantee a word-for-word performance.

Step 4 — One take, lips and voice together

The video model receives the approved script, the creator reference and the brand assets. For a selected creator, the prompt aims for one continuous take with a consistent presenter and setting. Other generation options can use a sequence of actions or shots.

Speech and mouth movement are generated together for this format. The presenter is synthetic, and the result still needs a human check for speech accuracy and natural lip sync.

The prompt to the video model is strict about what it may not do: no split screens or collages, no burned-in captions or fake UI text, no second location, and — since the script already fits the clip — no rushing the delivery.

Step 5 — Into the editor

The finished clip opens in Advinky's editor, where you can trim, add captions in Instagram-style presets, lay music under the speech, add an end card, and export. Captions are transcribed from the generated audio. Check their wording and timing before exporting.

From there, publishing to Meta is one click, and the approval step applies there too: nothing goes live without a person choosing to publish it.

What it costs

UGC videos are metered per second of rendered video at the chosen resolution. Review the estimate before generation. Completed work may consume credits even if a later step fails or you cancel. Contact support about incorrectly consumed credits; payment refunds follow the Refund & Cancellation Policy. The credits article has the current figures.

Make your first UGC ad