A mobile-first sales tracker for African small businesses — record a sale in under 10 seconds, parse WhatsApp orders with AI.
The challenge.
Ten million small businesses in Nigeria alone — and tens of millions more across Africa — run their sales on paper scraps and WhatsApp threads. They don't know their profit, they lose records, and they can't get loans because they have nothing to show a bank.
The constraints are real: cheap Android phones, patchy networks, noisy markets, sub-second decisions between customers. The product has to disappear into the workflow.
The solution.
Mobile-first PWA on Next.js with 48px tap targets, offline-first caching, and a sale flow that remembers your last product, quantity, and payment method — most sales are repeats, so the defaults do the typing.
The differentiator is the WhatsApp parser: paste an entire customer message, and a pattern-matching engine extracts items, quantities, totals, and payment method, fuzzy-matching against the user's product catalog ("Coke" → "Coca Cola 50cl"). Confidence score surfaces uncertain parses for review. Backend is NestJS + Postgres with JWT auth and strict per-business data isolation. Cost-price tracking turns every sale into a profit number on the dashboard, not just revenue.
The impact.
Sale entry compressed from ~45 seconds on paper to under 10 seconds in-app, and ~15 seconds for a parsed WhatsApp order. Across 100 sales a day that's roughly 3.5 working days a month returned to the operator. Sellers using the beta report real profit visibility for the first time — and the records to back a loan application.