Learning Timeline
Key Insights
Audio is Key
Strong audio is far more important than perfect visuals. Viewers usually won't notice minor facial defects (visual artifacts) if the audio storytelling is compelling and fast-paced.
The 5% Hit Rate Strategy
Don't settle for just 2 or 3 generations. Professional creators generate up to 300 clips just to get 15 usable clips. The quantity of generation directly determines the quality of the final output.
The Advantage of Public Tools
Viral ads don't necessarily require secret tools. This Kalshi ad was produced using publicly available tools that anyone can access right now.
Prompts
Street Interview Visual Generation
Target:
Google Veo
A hyper-realistic close-up shot of a person on a busy Florida street, talking excitedly to the camera, authentic facial expressions, cinematic lighting, 4k, handheld camera movement style.
Step by Step
Kalshi-Style Viral AI Ad Production Workflow
- Draft a clear 'Vision' and ad script before starting any AI generation.
- Use ElevenLabs to generate voiceovers or 'street interview' style dialogue with realistic emotion and intonation.
- Access Google Veo (or equivalent AI video tools) to generate visuals based on the script scenes.
- Input specific prompts for each scene, focusing on human movement and facial expressions.
- Perform bulk generation. Generate at least 300 video clips to ensure you have enough variety.
- Filter the results and select only the highest quality and most relevant clips (aim for a 5% success rate or the 15 best clips).
- Import the ElevenLabs audio and selected video clips into video editing software (such as CapCut or Premiere Pro).
- Arrange clips using 'fast-cutting' techniques synced to the audio rhythm to maintain momentum and mask AI visual artifacts.
- Add a layer of upbeat background music to reinforce the 'unhinged' or viral vibe of the video.
- Perform the final export in a format suitable for social media (9:16 for TikTok/Reels).
