ba2olak: The Delivery Marketplace We Built for Egypt's Underserved Towns

Grocery and market delivery in Egypt has scaled fast in Cairo and Alexandria — but ask anyone living outside the major cities and you'll hear the same thing: none of the big delivery apps reach them. We set out to build a delivery marketplace specifically for the towns and neighborhoods that the existing players have left behind, launching it first in the areas where the need was clearest.

The Problem: Delivery That Never Reaches These Neighborhoods

Grocery and market delivery in Egypt is concentrated in dense urban areas. Established delivery apps rarely reach smaller towns and remote neighborhoods, so residents there still walk between multiple stores, queue, and carry their purchases home themselves — the exact friction that delivery is supposed to remove, still fully present for a large share of the country.

What We Built

ba2olak is a bilingual, Arabic-first web and mobile marketplace. A customer orders grocery and market items from their phone; a ba2olak driver shops for those exact items in real, local stores and delivers them to the door, cash on delivery. The product is deliberately built around the constraints of a new market rather than a scaled-down copy of a big-city app.

• One Expo app, two roles — the same app serves both the customer ordering flow and the driver fulfillment flow, so onboarding a new driver doesn't require a separate build.
• A crowd-sourced, AI-deduplicated catalog — instead of manually digitizing every store's inventory before launch, the catalog grows from what drivers and customers actually enter, with an AI pass catching duplicate and near-duplicate items so it stays clean as it scales.
• WhatsApp OTP authentication — no email/password form to fill in; verification happens over the messaging app almost everyone already has open.
• An admin dashboard for dispatch and catalog moderation — giving the team a single place to assign orders and keep the crowd-sourced catalog trustworthy.
• Bilingual from the ground up — Arabic is the primary language throughout the product, not a translated layer added after the English version shipped.

Design Decisions Worth Explaining

A few choices weren't obvious upfront. Making cash on delivery the default rather than an option meant rethinking how driver settlement and reconciliation work, since the money doesn't move digitally until the very last step. Letting one person be both shopper and driver kept the platform viable at low order volumes in a brand-new area — a dedicated shopper role only makes sense once demand is proven. And crowd-sourcing the catalog instead of building it centrally traded a slower, messier start for the ability to launch in a new town without months of manual data entry first — the AI deduplication pass exists specifically to make that trade-off safe.

Where Things Stand

Lessons We're Taking Into the Next Build

1. Cash-first payment design has to be planned from the data model up, not retrofitted — driver settlement logic is much harder to bolt on afterward.
2. A single flexible role (shopper + driver) is the right default for any new-market launch; specialize only once volume justifies it.
3. Crowd-sourced data needs an automated cleanup pass from day one, or the catalog degrades faster than a small team can moderate it by hand.

If you're weighing whether a similar approach fits your own underserved market, the perfect delivery experience breaks down the decisions in more general terms, and the ba2olak case study has the full project detail.