🚨🇺🇸 Google a ajouté Photoshoot à Pomelli, une nouvelle fonctionnalité IA qui permet de créer des supports marketing sans avoir besoin d’un studio professionnel
Ajouter un sondage post-achat est une MINE D’OR !
Pour comprendre quoi améliorer concrètement afin de :
> créer de meilleures ads
> augmenter le CVR
> repositionner la marque intelligemment
Les réponses clients te disent quoi corriger,
au lieu de deviner.
Voici les top 4 questions que j'aime poser lors du thank you:
- Quelle était votre première impression de VOTRE MARQUE ?
- Comment avez-vous entendu parler de nous ?
- Avez-vous rencontré des difficultés lors de cette commande ?
- Y a-t-il eu quelque chose qui a failli vous empêcher de finaliser votre achat ?
$60k+/month ad accounts are already leaning into animated objects
Not mascots.
Not cartoons.
Products that behave like characters inside short-form videos
A lemon reacting to heat
An onion getting angry
A leaf calmly sitting in water
It feels like content, not advertising
That’s the entire advantage
People have learned to scroll past faces
They haven’t learned to scroll past objects doing something unexpected
AI generates the characters
AI generates the scenes
AI generates variations
Same idea
Hundreds of executions
No filming
No creators
No scheduling headaches
Once one version works
You multiply it
This format compounds because it’s simple, weird, and scalable
rt + comment “objects” and i’ll dm the setup
(follow for dm)
and it only takes a few minutes to create
simply:
- nano banana pro for first frame
- veo to animate it
- capcut for editing
done, a skill that’s in huge demand is just 3 simple steps and a bit of aesthetic sense
wan 2.2 is starting to feel borderline unfair
you start from one talking head clip
drop in a single face reference
and the model reconstructs the entire video
same delivery
same timing
same emotional cadence
completely different person
brands are using this to clone winning ads
across markets
languages
and audiences
without ever filming again
one video turns into a full distribution engine
i put together a short breakdown that covers
how to choose source clips that survive face swaps
how to keep the outputs realistic
and how to generate variants that still convert
rt + comment “wan” and i’ll send it
(follow for dm)
Going to regret leaking this...
I made a step by step cheatsheet on how to make these realistic AI videos for your ads
Like, RT & comment "Video" and i'll send it over
(must be following)