I Cracked Google's New Omni Flash to Make $100K Ads (Any Brand)

What this lesson covers

Adrian demonstrates how to use Google’s Omni Flash model to create realistic AI‑generated UGC and animated ads, covering image creation, upscaling, voice integration, and lip‑sync quality. He also shares workflow tips using Claude prompts and a Flow Agent to streamline production for high‑impact advertising.

Key takeaways from the creator

AI-extracted notes, not independently verified product claims. Timestamp links let you check each point in the original video.

1. 02:00 ↗

   Upscaling from 720p introduces plasticky artifacts, especially in teeth, highlighting current resolution limits of Omni Flash.

2. 05:45 ↗

   Adding a voice after setting the start frame reduces the model’s visual fidelity to that frame, so voice should be specified in the prompt instead of as a separate ingredient.

3. 07:14 ↗

   The lip‑sync and camera movements produced by Omni Flash are markedly better than the older VO3 model, making it suitable for realistic talking‑head UGC.

4. 07:56 ↗

   Successful realistic UGC hinges on precise prompts and high‑quality source images; framing subjects close to the camera preserves detail during upscaling.

5. 08:01 ↗

   A simple workflow—prompt generation in Claude, image selection, import into Google Flow, then animate with Omni Flash—enables rapid ad creation.

Workflow outlined in the video

1. Write detailed Claude prompts that include desired voice characteristics before image generation.
2. Generate multiple high‑resolution source images, keeping subjects framed tightly to capture facial detail.
3. Import selected images into Google Flow as ingredients and set them as the start frame for animation.
4. Run the Omni Flash model to produce the video, then upscale only if necessary, checking for artifacting on faces and teeth.
5. If a voice is needed, embed voice instructions in the original prompt rather than adding a separate voice ingredient.
6. Review the final clip for lip‑sync accuracy and camera motion; re‑run the model with adjusted prompts if needed.

Before you use this workflow

These notes describe the source video at its publication date. Model access, pricing, connectors and interfaces may have changed. Check the original source and the provider’s current documentation before installing an add-on, connecting an account or spending credits.

ReelStack has not independently tested this workflow. Preview one representative shot and check motion, continuity and output quality before applying it to a full production. No result, cost saving or model capability is guaranteed.

Explore the tools

How to interpret the numbers

4,493 views captured 24 October 2026. ReelStack recommendation score: 34/100. Formula: radar-v5. This is ReelStack’s calculation, not a YouTube rating or a measure of factual accuracy.

Read our methodology and limitations →