A Finishing Story
Client: Omelet
Role: Lead 2D Motion Design , Edit Refinement
Background
To get people upgrading to Google AI Pro, Omelet built the campaign with the product: DR ads generated through Google's own AI stack, Gemini, Flow, Whisk, and more. Charming characters, eye-popping worlds, machine-made. An AI artist generated the material and laid out the edits.
I came in downstream: refining the cuts and building the motion design layer, the human-made finish on AI-made footage.
Problem
AI-generated footage doesn't behave like shot footage. Clips were never filmed to match, so the seams between them want to show. Cuts that would feel natural on real coverage read as jumps. And generated material can't make graphics at all: the UI, the type, the motion design that turns output into a Google ad still had to be built by hand.
Challenge
Make the seams disappear and the brand appear. Frame-level work: trimming each shot until the cuts felt smooth instead of stitched, then animating the graphic layer over footage with no camera behind it, to Google's polish bar.
Approach
I refined the edits frame by frame, cutting into each shot until the transitions read as intentional, the difference between AI clips in a row and a spot. Then I built the motion graphics and UI animation across the campaign: the type, the interface moments, the layer that makes the spots read as Google rather than as generations.
Working downstream of an AI pipeline taught me exactly what machine-made footage needs from a motion designer, and where the human layer still does the convincing.
Results
The campaign shipped 126 assets, 68% faster and 60% cheaper than Google's previous DR campaigns, over 2,000 prompts behind the generated material. My motion layer and cut refinement are part of what carried that output to broadcast-ready.