About
Health questions do not wait for the appointment.
Reports kept arriving for my wife and my father that none of us could properly read — and the AI everyone already had on their phone was confidently wrong about them.
Here, the doctor does not hand you your results. The doctor says get this test done, you go to a laboratory, and a few days later a PDF arrives on your phone. It is precise, it is complete, and it is written for the doctor who ordered it. The appointment is weeks away. In between there is nothing.
So we did what everybody does — we typed the values into a general-purpose chatbot, the kind already sitting on every phone. Sometimes it helped. Often it was confidently wrong. It answered against reference ranges from somewhere else entirely, it restated numbers that were not on the page, and it never once said it did not know. Asking about my wife’s results, it would reassure or alarm with exactly the same certainty, and with no source behind either.
That is a specific failure, and it is not really the chatbot’s fault. A general model is built to answer anything. Health is the one subject where answering anything is the wrong instinct — where the useful answer is usually narrow: this is what the value is, this is the range your own laboratory printed, and this is the question worth asking about it.
So we built something that does only that, and does it in the language the person actually thinks in — 98 of them, because the people who most need this are rarely served in their first one.
What we are
A healthcare AI company
We built our own model for health, and the app is the first thing we put behind it — not the whole of it. The same model works for organisations: an agent that takes the administrative work off a clinic so the doctor consults and signs, and custom systems built into what a laboratory, a pharmacy or a manufacturer already runs.
What it does is the part worth stating plainly. It reads a clinical document and explains it against the ranges printed on that document. It answers in the language it was asked in. It says so when it does not know. Those limits are the product, not a caveat on it.
What we build for organisationsWhat we hold ourselves to
Four rules we do not trade away
Every one of these costs us something — a feature, a shortcut, or a revenue line. That is rather the point.
We read your lab's range, never our own
Sourced, or not said
Explain, never diagnose
Your data is not the business model
What we are not building
Worth saying plainly, because each of these would be an obvious way to make money and each would make the product worse.
- We are not a route to a doctor. We do not refer, book, or take a fee for sending you anywhere.
- We are not a diagnostic tool, and we are not a medical device.
- We do not sell health data, and we will not make your results available for insurance or eligibility decisions.
- We do not tell you what to eat because of a value on your blood test.
Ammar Naqvi