Best Practices for Voice Video Ads

A Dialogue Video Ad is made up of the following components:

  1. 1.
    Teaser video, up to 15 sec
  2. 2.
    AI response if interested
  3. 3.
    AI response if not interested
  4. 4.
    AI response if silence
  5. 5.
    AI response when listener’s response doesn’t make sense
  6. 6.
    Additional AI response if the listener provides another response that doesn’t make sense
  7. 7.
    AI responses if the user responded with a swearing word (Optional but recommended) .
  8. 8.
    Custom AI responses if you need a specific reaction for a certain phrase.

Technical requirements

  • Please make sure you have no silence at the end of the teaser and at the beginning of all AI responses.
  • The best format is mp4. It’s better to prepare the videos in 1080p.
  • Make two versions of all the videos: one vertical and one horizontal.
  • If you need to show a picture during the transition between videos, add the desired picture as the last frame of the teaser video and as the first frame of all the AI responses.
  • If you need to show subtitles, prepare a WebVTT subtitles file for each video.
  • Please keep the volume level normalized to -10db with occasional peaks possible at -8db.