Playable uses previous video campaign data to train its AI on what works best, and then uses this AI to automatically edit your next video campaigns for optimal engagement.

The video opposite provides a glimpse of this frame by frame analysis in action where computer vision is being used to watch every pixel of every frame, to identify the objects, events, faces, and sentiments of a video.

Different predictive models are created for different success metrics, depending on your marketing objectives. Typical objectives against which models are developed include viewing duration, click-through rate, number of shares, or downstream conversion metrics, such as basket or checkout events.

The technology has been built in conjunction with a videographer with 100’s of elements of the craft captured.

The AI then uses the data from previous video campaigns to generate new videos. Theoretically, it can make an infinite amount but it starts with 10 – 50. Some edits are shorter, some are faster, some focus on people, some on objects.

The AI also considers the importance of brand and messaging, with the inclusion of branding and key messaging within the video either from content within the source video, or by adding logos and text overlays.