Why we build the rails first

People ask why a company that cares about accurate health information spends so much of its time on identity registries and interoperability APIs. The honest answer is that you cannot have one without the other.

The gap is underneath

Consider what a good answer to “is this drug safe for me?” actually requires. It requires knowing who is asking — their age, their genotype, what they are already taking, what happened at their last visit. It requires that information to exist somewhere, in a form that travels. And it requires a model that can reason over it in the language the person actually speaks.

In Nigeria today, almost none of that exists. There is no shared health identity. Records live on paper that does not leave the building. Labs, pharmacies, clinics and insurers cannot see each other. The best model in the world, dropped into that environment, is answering blind.

So a “health chatbot” that is merely a good model is a demo. It will be confidently wrong in exactly the cases that matter most.

The pattern

Kindred’s answer is a sequence, and the sequence is the company:

1. Identity. A universal health ID that belongs to the person and works offline.
2. Records. A data layer that reasons across structured records, conversation, images and reported symptoms — not just the fields a 1990s EMR expected.
3. Rails. Interoperable APIs connecting providers, labs, pharmacies and insurers, so the record actually moves.
4. Reasoning. Intelligence built into the system itself, from our own models and fine-tuned open weights, shaped to African demographics, languages and practice.

Each layer makes the next one trustworthy. Accuracy is what you get when all four are in place. It is not a feature you can add to a prompt.

Why health first

Because the gap is widest and the stakes are highest. And because health forces discipline: if we can be accurate here, under these constraints, with a doctor signing off on every safety standard, we can be accurate anywhere.

Why it compounds

Once a patient has an ID, the provider tools become worth using. Once a clinic runs its day through the platform, the records deepen. Over time, Alma knows a person’s medications, trends and history in a way nothing bolted on later can replicate. That is not lock-in for its own sake. It is what accurate information looks like over time.

What this is not

It is not an app company. Energy, agriculture, education and finance have the same shape: a missing identity, a missing shared layer, and intelligence that could reason over both if they existed. Health is where we prove the pattern. The rails are the point.