AI at Work

AI-generated user content ads cost five dollars versus five hundred

A demonstrated workflow for creating AI avatar ads reduces production costs to approximately $5 per video, compared to over $500 for human-made content. The process enables high-volume creative testing, which Schneider asserts is essential for modern ad algorithms.

· 1 source · Ad production costs

A specific workflow for creating AI avatar user-generated content ads has been demonstrated, costing approximately $5 per video compared to more than $500 for human production. The process involves writing a script, generating audio via FAL voice cloning, and sending the audio, script, and reference image to Seed Dance to create a 480p video. This video is then upscaled and blurred using FFmpeg to improve quality. Word-by-word transcription is used to detect and correct pronunciation errors, such as for the word "Calendly," by swapping audio segments. Motion graphics and captions are added using Tesseract.

The workflow was tested on brands including Idea Browser, LateCheckout, Cursor, Robin Hood, and Teenage Engineering, generating static ads and video concepts based on scraped brand styles and Reddit pain points. Human creators can also be integrated via APIs from Tribe or SideShift, paying approximately $40 per video. Schneider asserts that AI avatars can sometimes outperform human creators in ad performance. He argues that the low cost enables high-volume creative testing, or "creative velocity," which is essential for modern ad algorithms. Schneider noted, "I can go and pay $4. Yeah. 95% is good out the other side."