Every company can make content now. That is exactly the problem. When anyone can produce anything, volume stops being an advantage — and the only asset you hold that a competitor cannot copy is your own story. We turn it into something people choose to watch.
CanEl is an AI studio built by storytellers. Not a marketing company that adopted AI — a company built with it from the ground up — in choosing what to make, not just in making it.
Anyone can prompt a model. We built an operating system around it — where decades of making television govern what gets asked for, and what comes back. Every production leaves three things behind.
Characters, locations, sets, wardrobe, lighting. Starting states for the next piece, not finished work.
Workflows, continuity, model configurations — and the failure data, which matters more.
Every output, the one chosen, the ones rejected, and why. Human judgment, on the record.
We are not tied to any AI vendor. The models are rented equipment — when a better one ships, we switch to it, and nothing about how we work changes. See how it works →
That is what the process runs on, and what it feeds back into.
Reach is purchasable. Belief is not. An impression is rented and forgotten; a story someone chose to watch, and felt something during, leaves them believing you understand something they care about.
A financial firm’s thesis: investors who believe they missed a rally are usually watching the wrong indicator. Two ways to say it.
Well made. Competing with everything else in the feed.
A man runs into a station and finds every board reading DEPARTED — until someone shows him a second board, where price flattens and earnings keep climbing. Thirty-nine shots. It never states its argument out loud.
Same idea. Different medium. Different result.
Client work is shown without attribution. Named references and full case studies are available under NDA — request them.
Two pieces built on the same system, in a different register. Both produced start to finish the way your work would be.
Where the machine works and where we do. The eighty in the middle is the part everyone talks about. The twenty on either side is the part that decides whether it is any good.
Structure, rules, constraints, and worked examples down to the sentence. The system only produces what it was told to, and telling it is the job.
Bulk production, inside the rules we set. Fast, tireless, and completely dependent on the ten percent that came before it.
Tone, nuance, word choice, rhythm. The difference between raw output and something a person actually wants to sit through.
Written and broken into numbered shots before a single image exists.
Built as a full sheet, then locked as the reference every later shot inherits.
Nothing moves until the still frames are signed off and locked. No video is rendered, and no budget spent, on a look nobody has approved.
Lighting, lens, grade and grain set once, applied across the run.
Generated against the locked references, then cut and finished like any other film.
Two partners, one conviction: the story is the product.
Jeremy developed premium television at AMC across the run that defined the network — Breaking Bad, Mad Men, and The Walking Dead — winning two Peabody Awards, for Mad Men and Breaking Bad. He built the television department at Legendary Entertainment from scratch and held a production deal at Blumhouse. He has seven published novels.
Sean has spent fifteen years across CBS Studios, Electus, Propagate Content, and McFarlane Entertainment, developing series for Netflix, Prime Video, Paramount+, CBS, and The CW. His credits include the Peabody-winning Jane the Virgin, Blood & Treasure, No Tomorrow, and Flaked.
Not commentary. Primary data from work we actually delivered — real numbers, real failures, and the parts nobody selling AI tools will tell you. Published here first.
Everything we publish lives here first. The full archive is on this site; the newsletter is how it reaches your inbox.
Read all Field NotesStart here
Client work, partnerships, original IP, investment, and conversations about where any of this is going. We read everything, and we answer.
