MOTION 02 +
What's Your Smile Worth To You?
2026
ABOUT
An Ai driven emotional Ad aimed at making the viewer think about why they might be hesitant over getting treatment. To be ran as an ad for Meta and also posted as a Short for Youtube.
Editing / Ai Generation
Final film +
Best watched with sound on
01 - Concept and Brief +
CONCEPT &
BRIEF
Client
Length
Format
Role
21D Clinical
1 min
16:9, 9:16, 1:1
Project Lead & Editor
THE BRIEF
Use Higgsfield to create Ai videos and use these to edit an advert for use in socials and campaigns.
A company decision to embrace Ai video generation to take off some of the workload and reduce production time and cost.
​
I was tasked with creating an emotional advert aimed at covering patient self-worth and relatability.
​
With the provided script I was to create an engaging and emotive narrative for use in Advertising as part of a branding campaign.
02 - Toolkit +
TOOLKIT RUNDOWN
Three main tools were used for this video creation.
Tool 01 / AI video
Hf
Higgsfield
Generated the footage for the ad from written prompts, shot by shot.
Video generation
Tool 02 / AI voice
EL
ElevenLabs
Generated the voiceover from the original voiceover and exported it as a more appropriate clean voice.
Speech generation
Tool 03 / Adobe
Pr
Premiere Pro
The main edit: Used for putting together final version, adding sound effects and music and preparing for export.
03 - Workflow breakdown +
WORKFLOW BREAKDOWN
A comprehensive breakdown of the ai workflow which the team pioneered, adapting current methods of work to embrace the new tool.
01 - Plan shots from script
Record a few readthroughs of the script to structure length and pacing.
Seperate the voiceover into separate lines to use as script for each given scene.
Replicated Prompt for reference.
02 - Claude + Higgsfield
The main workflow was to take an idea and write it up in brief. Then feed this brief into Claude to give back to us as a detailed prompt.
We would then utilise the canvas feature of higgsfield to generate a properly structured storyboard.
We would start by generating images as these used very few credits and then consistently feed back to claude, asking for tweaks and changes until the generation felt as the vision intended.
After this was solidified we would feed the finalised prompt into higgsfield, mainly opting for Seedance 2.5, until we got a useable generation.
This process repeated until all scenes were complete, ensuring consistency between the shots. Sometimes we then went back through to gain additional angles of specific shots.
The Canvas feature also allowed the entire team to collaborate and pick up the slack whenever a project fell behind.
02.1 - Higgsfield Canvas +
03 - Premiere Pro assembly
After the shots are finalised and downloaded from higgsfield they were imported into Premiere.
Here the editing was largely the same as any other project.
The advantage was the ability to easily "reshoot" (generate) any additional shots which were needed for the edit on the fly.
04 - ElevenLabs speech generation
For these types of videos I opted to use a different method of voiceover.
Instead of the standard line by line generation, I fed my own voiceover into ElevenLabs and converted it into the desired voice.
This allowed the appropriate inflection on letters and words to be captured without multiple generations.
05 - Sound design
After the voiceover was finalised and placed on the edit, I used stems taken from Epidemic Sound to get the correct feel of tension for the film.
I then layered in passes of EQ and additional sound effects for extra detail.
06 - Back to Premiere Pro
After all this was layered together, I did a final pass in Premiere for last tweaks and internal review, before exporting via Media Encoder.
After approval I then went back into Higgsfield to create the 16:9 and 1:1 aspect ratio generations, using the original 3:4 as reference inputs.
