Your five steps, one by one
Each step gets its own beat and one large, simple shape, so nobody has to read anything to follow it.
I built a first draft of your break-room clip to show how I would handle it: one employee, your five steps in order, no text, no sound. The phone screen is real motion graphics laid over the 3D footage, so every icon stays sharp from across the room.
Best seen full screen from a few metres away, the way it will run on the cafeteria TV.
Each step gets its own beat and one large, simple shape, so nobody has to read anything to follow it.
AI where it looks good, classic compositing where precision matters.
Semi-realistic 3D stills of the employee, her hands and the break room, generated with Nano Banana Pro and kept consistent from shot to shot (same face, sweater, skin tone).
The phone is filmed with a green screen. Camera icon, scan frame, link button and answer faces are drawn as vector graphics in Remotion and placed on it, so they never blur, warp or show random letters.
The tapping hand is a separate layer moved by keyframes: it slows down, presses (the icon shrinks a little, a ring spreads) and lifts. Every tap lands on the exact spot.
The opening shot gets real AI motion (image-to-video). The final version uses whichever video model handles faces and hands best for that shot, checked frame by frame.
Every frame is checked for letters, numbers, logos and fake writing, the phone screen included.
Delivered without an audio track, so the TV never plays a sound.
210 frames at 30 fps = 7.0 seconds, in 1920 × 1080 or 4K if you prefer.
Fingers counted, taps checked, the employee looks the same in every shot.
With your creative brief and reference video I adapt the look (colours, workplace, clothing, the answer icons and the final shape), send a preliminary version for review and work your feedback in over up to three rounds. You get the final MP4 without any branding or audio, with full commercial rights.