This week I want to hand you one story and get out of the way.
She spent 17 years at IMAX. Then AI let her build a studio of one.
Kelly Boesch did graphic design, film production, and marketing at IMAX for 17 years — building the development decks that convinced studios to fund projects. Three years ago she picked up AI image generation. She didn't dabble.
She now has 2–3 million followers, around 60 million views a month, a debut album on Nettwerk Music Group, and a recent gallery show in London. The trajectory is eye-catching. But it's the process underneath it that I keep wanting to understand — so we wrote up how she actually works.
The thing that reframed it for me: most people approach AI video text-first — describe the scene, generate, hope. Boesch works the opposite way. She generates a still in Midjourney, gets the composition exactly right, then animates from that image. Text-to-video is closer to a lottery. Image-to-video is closer to art direction.
Then she writes her own lyrics and directs Suno to produce the music — which is how a cohesive 14-track album ends up on a real label the same week as her London show.
The full story gets into:
- Her actual toolset and why she moves between Runway, Pika, Higgsfield, and VEO3 depending on what a scene needs
- Why the edit layer — color correction, pacing, sound mixing — is where she says the human touch turns a clip into a story
- What 17 years of professional craft gives her that the tools can't, and why that's the part worth studying if you're starting now
If it lands, hit reply and tell me which part of her process surprised you. I read everything that comes back here.
— William