Glossary · Making it with AI
Image-to-video
Image-to-video is a kind of AI that brings a still picture to life, using it as the starting frame and adding the motion a written note describes.
Why it matters
Because the picture fixes how everything looks, the AI only has to invent the movement. That makes the result more predictable than starting from words alone, and keeps people and places looking the same across shots.
The still picture sets the people, the setting, the colors and the framing. The written note then describes only what changes: who moves, how fast, and what the camera does. Describing again what the picture already shows tends to work against it.
Some systems also accept a last picture, so both the start and the end of a shot are set and the AI fills in the motion between them, much like keyframes in animation.
It still needs checking. Large movements can warp faces and hands, and anything the picture doesn't show, like the far side of a room, has to be invented. Simple movement and short shots reduce those problems.
An approved picture of a young engineer at a drafting table becomes a five-second shot: she looks up from the plans, the desk lamp flickers, and the camera pushes in slowly.
How we use it
Every moving shot in our videos starts as a still picture, approved on the storyboard, then brought to life.
Last reviewed 15 October 2026.