the context.
Olilo produces AI avatar videos for companies. To show the level the process reaches, Leandro Valino, Olilo’s founder, became the model himself.
the challenge.
A generic avatar gives itself away as AI fast: hands lose detail, an accessory changes from one frame to the next, the setting does not exist. Here, people who know Leandro had to recognize his watch, bracelets, wedding ring and the room where he works.
what we did.
We built the avatar on five fronts:
- Model sheet: front, profile and back portraits, full body from four angles and both hands, with watch, ring and bracelets.
- Real setting: we photographed Leandro’s office and used the photo as reference. The avatar appears in his room, with the same signs, the guitar and the neon.
- Three looks: the same avatar got three combinations of outfit and setting, to vary the videos with no new shoot.
- Voice and script: the text is written and approved first. Only then does the avatar record, in Leandro’s voice.
- Two formats: the same video came out in horizontal and vertical, with captions.
result.
The video runs 32 seconds and opens with the avatar announcing, in Portuguese: “this one here is not me”. Watch, bracelets, ring and even the iPhone wallpaper look the way they do in real life. From here on, each new video starts from an approved script, with no shooting day.
who it is for.
The same process works for business owners and content creators who need to publish often and have no room in the calendar to record. You approve the script and receive the finished video.
