The drugs worked. That was the hard part.
Incretin therapies produce weight reduction previously achievable only through surgery. What surrounds them is a twenty-minute visit every few months and a printed handout. We are building the part in between.
Our customer’s business changed this year.
In roughly sixteen months the molecule became a commodity with a published price, public coverage began, and the compounded-supply margin was enforced out of existence. Every dollar that used to sit in the medicine has left it.
What remains for a weight-management practice is the care around it, and the only product it has left to sell is whether patients stay. That is not a marketing problem. It is an operations problem that happens in the weeks nobody is watching.
Gallup’s National Health and Well-Being Index put current GLP-1 use for weight loss at 11% of US adults in June 2026, up from 3% in 2024.
It cannot be staffed.
The obvious answer is to hire someone. The numbers say no, and they do not say it narrowly. A registered nurse costs about $101,000 a year in wages alone before benefits or overhead. Between-visit work for a panel of several hundred patients is more than one person’s job, and unlike a visit, none of it bills.
So it is the first thing a practice cuts and the last thing it can afford to. Which is why this care does not happen today, and why when it does happen it will be executed by software with clinical judgment reserved and routed rather than replaced.
US Bureau of Labor Statistics, May 2025, mean annual wage for registered nurses. Per-patient-per-month figures circulating in this category generally trace to vendor marketing rather than to a source, so we do not use them.
Everyone who could build it is disqualified.
The components are unremarkable: scheduled messaging, structured capture, classification, routing. The question is why, in a category this large, nobody has assembled them. Each class of incumbent is ruled out by its own structure, and none of those reasons is going away.
The honest counter-argument is that between-visit engagement has been attempted repeatedly and failed on demand; twenty years of disease management disappointment is real evidence for it. Our answer is that the buyer, the cost per interaction and the absence of a formalized program object have all changed at once, and that this is cheaply testable rather than something to argue about.
Three commitments that decide the rest.
Small, and hiring for the seats that matter.
Jeni is early by design. Byungsoo Ko built both halves: the consumer weight product whose paying subscribers produced the insight and supply the patient-side design capability, and the clinic platform, which is being built with its first practices rather than sold ahead of them. If you book twenty minutes, that is who you are talking to.
We are assembling a founding team around three seats: clinical, engineering, and go-to-market. The clinical seat matters most and comes first: protocol authorship is the durable asset, and a company where authorship depends on one or two people grows its protocol library linearly rather than compounding.
If you are an obesity medicine physician who has run this work by hand and knows exactly where it breaks, or an engineer who wants the hard part to be correctness rather than scale, write to us at hello@jeni.health.
The layer gets built by someone.
Better that it is built carefully, with the clinics who will run it, and measured in public.