Client project · Wholesale e-commerce
Islamic Hijab
A wholesale store for modest fashion, selling to shops across Europe and Africa. Its suppliers post new pieces on Telegram all day. Our AI pipeline turns each post into a ready store product, and the team only reviews and approves.
- Client
- Islamic Hijab
- Industry
- Wholesale modest fashion
- Market
- Shops in Europe and Africa
- What we built
- Wholesale store + AI product pipeline + admin system
- Our role
- Discovery, design, frontend, backend, AI pipeline, launch and support
The challenge
Almost every step was done by hand
Islamic Hijab buys from many suppliers and sells in bulk to shops abroad. Before the new system, almost every step was done by hand.
- New products arrived as Telegram posts.
Suppliers publish dozens of new items a day in their channels, with photos and messy text: sizes, colors, pack size and price, often mixed together.
- Every product was retyped.
Someone saved the photos, typed the name, sizes and colors, and uploaded the product, one by one. It took hours, and new items reached the store late.
- Photos didn't look like one brand.
Each supplier's photos had their own style. Adding the store's logo and design was another manual job, so many photos went up without it.
- Prices were worked out by hand.
The team added a profit margin to each supplier price with a calculator. Different categories need different margins, and mistakes were easy to make.
- Sold-out items stayed online.
Fashion stock moves fast. Products the supplier no longer had stayed in the store, and customers ordered items that couldn't be shipped.
- Wholesale pricing lived in chats.
Buyers who order more pieces expect a lower price. Every discount was agreed over WhatsApp, so prices were different for every buyer and hard to track.
- Sizes made orders confusing.
One design comes in many sizes. Buyers sent size lists in messages, and the team matched them by hand.
- Shipping abroad was expensive.
Buyers placed several small orders, and each one was shipped and paid for on its own, so shipping often cost more than it should.
- Orders were spread across chats.
There was no single place to see each buyer's orders, what was paid, and what was shipped.
Our solution
A wholesale store with an AI pipeline behind it
We built a complete wholesale store with an AI pipeline behind it. The pipeline follows the suppliers' Telegram channels, reads every new post, and builds a draft product with the store's design, logo and prices. Drafts wait in a review area, where the team checks, edits and approves them. Approved products go live in the right category.
On the store side, everything is built for how wholesale really works: prices that drop with quantity, many sizes per design, products that hide when their time is up, and shipments that combine several orders into one cheaper delivery.
Problems we solved
Every problem, and what we built for it
| The problem | What we built |
|---|---|
| New products arrived as Telegram posts | A pipeline that follows supplier channels and picks up every new post automatically |
| Every product was retyped | AI reads each post and fills in the name, category, sizes, colors, pack size and supplier price |
| Photos didn't look like one brand | Every photo gets the store's design and logo automatically |
| Prices were worked out by hand | Profit margin rules per category, applied to every product |
| AI drafts need a human check | A staging area where the team reviews, edits, approves or rejects each product |
| Sold-out items stayed online | Category timers: products hide automatically after the time set for their category |
| Wholesale pricing lived in chats | Tiered pricing: the price per piece drops as the quantity goes up, the same for every buyer |
| Sizes made orders confusing | Multiple sizes per product, chosen with quantities in one order |
| Shipping abroad was expensive | Shipments are separate from orders: buyers combine several orders into one cheaper shipment |
| Orders were spread across chats | Buyer accounts and an admin system with every order, payment and shipment in one place |
Key features
What the system does
AI product pipeline
Supplier posts in, ready store products out. Nothing is published without a person's approval.
Your look on every photo
The store's design and logo are applied to every product photo automatically.
Margin rules
A profit margin per category is added to every supplier price, so prices are always correct.
Review area
Check AI drafts, fix anything, then approve or reject, one by one or in bulk.
Category timers
Each category has its own time limit. When it ends, products hide from the store on their own.
Tiered pricing
The more pieces a buyer orders, the lower the price per piece.
Multiple sizes
One product, many sizes, with quantities chosen per size.
Combined shipments
Shipments are separate from orders, so buyers group several orders into one shipment and pay for shipping once.
Highlight · The AI product pipeline
From a supplier's post to a store product, in minutes.
Suppliers post new pieces on Telegram all day. Our pipeline picks up each post as soon as it's published. AI reads the photos and the messy text and pulls out the details: name, category, sizes, colors, pack size and price. Each photo gets the store's design and logo, and the category's margin is added to the price. The finished draft waits in the staging area until someone on the team approves it. Then it goes live.
- Stop 1Supplier postsTelegramnew jersey hijab!! 12 colors, pack of 10, $2.40 eachTelegramAbaya Linen — sizes 52 54 56 58, pack 4, 14$
- Stop 2AI readsJersey hijab12 colors · pack 10 · $2.40Linen abaya52–58 · pack 4 · $14.00
- Stop 3Photos styledIHIH
- Stop 4Margin addedhijabs +35%
$2.40$3.25abayas +30%$14.00$18.20 - Stop 5Team reviewsJersey hijabEdited · ✓colors 12 → 11Linen abayaApproved ✓
- Stop 6Live in the storeJersey hijabLinen abayaHides in 14 days
- Stop 1Supplier postsTelegramnew jersey hijab!! 12 colors, pack of 10, $2.40 each
- Stop 2AI readsJersey hijab12 colors · pack 10 · $2.40
- Stop 3Photos styledIH
- Stop 4Margin added
$2.40$3.25 - Stop 5Team reviewsJersey hijabEdited · ✓
- Stop 6Live in the storeJersey hijab · hides in 14 days
- 01 Collect
- 02 Read
- 03 Style
- 04 Price
- 05 Review
- 06 Publish
Inside the system
A closer look
Mockups with invented data.
The review area: AI drafts wait for the team's approval. A product page: tier prices, sizes and the category timer. Build a shipment: three orders, one cheaper delivery.
Our role
Who did what
Ammar Gomaa
Product and UX, the store and admin interfaces, the review area.
Abdullah Zakaria
Backend, the Telegram connection, the AI pipeline, pricing rules and shipments.
Together
Discovery with the client, the spec, testing with real supplier posts, launch and support.
Results
What changed
- New supplier products reach the store in minutes instead of waiting in a queue to be typed.
- The team's job changed from retyping posts to reviewing and approving them.
- Every product has the same look and a correct price.
- Buyers see the same wholesale prices, pick sizes in one order and pay for shipping once.
- Sold-out stock leaves the store on time, without anyone remembering to remove it.
More projects
