Building tools for the health workers around me

Walk into a small pharmacy or clinic around Nsukka and watch the system: paper registers, WhatsApp groups for supplier orders, mental math for stock, memory for credit. It works — barely, heroically. Watching it work is my best education in what software should actually build.

What the observation taught

The workaround is the spec. The WhatsApp group isn't chaos — it's a supply chain that survived every tool that tried to replace it. Any software that doesn't understand why the group works (instant, universal, zero training) will lose to it. Build with the group's virtues, not against them.

Paper has virtues worth stealing. It never needs charging, works offline forever, and anyone can use it. Health-tech that ignores these virtues builds for demonstrations, not for the counter. The target is paper's reliability with software's memory.

Trust moves at the speed of relationships. Health workers adopt what a colleague they respect adopted — not what a pitch deck promised. Tools spread through counters, not conferences.

Interruption is the default. Real health workflows get interrupted every four minutes. Software that assumes focus (multi-step wizards, save-only-at-the-end) fails in the actual environment. State must survive interruption automatically.

What I'm building from it

Small, offline-first tools for the workflows I've watched strain: stock that counts itself, orders that ride the channels people already trust, records that survive power cuts. Modest problems, honestly understood. That's the whole strategy.