It started with an award for not overreaching.
Allayr began as a hackathon project. The version we built there won Most Responsible Use of AI — which turned out to be the thing worth building the company around.
The problem we kept hearing about
A Canadian clinic’s inbox isn’t one inbox. It’s a fax machine, a shared email account, and a phone that receives texts, all filling up independently while the people who need to act on them are seeing patients. The urgent item and the supply invoice arrive looking exactly alike, and the only thing sorting them is whoever has a minute.
That’s an organizing problem, not a medical one. It’s also the kind of problem software is genuinely good at — as long as the software knows where to stop.
Why the award mattered more than the win
At the hackathon, the obvious move was to build something that acted on its own. We built the opposite: a system that sorts, explains itself, and then waits for a person. That earned Most Responsible Use of AI, and it clarified what the product actually was.
In healthcare the interesting constraint isn’t what the model can do — it’s what it should be allowed to do without a person. We landed on: sort it, explain the reasoning, suggest a destination, and stop. Every product decision since has run through that.
Where we are now
Two of us, building. Allayr is pre-launch: no pilots, no paying clinics, nothing to show you in the way of results. We’d rather say that plainly than dress up a demo as traction.
We build in public — the decisions, the reversals, the things we got wrong — on theblog. If you’d rather shape it than watch it, theInner Circle is where that happens.
We answer our own email.
No SDR, no sequence. Leave your address and one of the two of us will write back.