The challenge
Two shops, two versions of the truth.
Each shop tracks its own stock on paper. Nobody can tell a customer in one shop whether the other has the item, so a sale becomes a phone call, a wait, and often a no. Reordering happens when a shelf looks empty rather than when the numbers say so.
The business is profitable and completely undocumented. Slow-moving stock sits for months before anyone notices, and the fast-moving items run out repeatedly because nobody sees the pattern until it has already happened.
- Sales lost because nobody could confirm stock in the other shop
- Reordering driven by how a shelf looks, not by what sold
- Capital tied up in slow stock that stays invisible for months
What we built
One inventory, visible everywhere, with the assistant on top of it.
The ERP first — the assistant and the site are only as good as the stock data behind them.
Intelligent ERP & CRM
One live inventory covering both shops and delivery. Stock moves as sales happen, transfers between shops are recorded, and slow-moving and low-stock items surface on their own rather than waiting to be noticed.
- One inventory across both shops and delivery
- Low-stock and slow-moving alerts, before they cost you
- Sales per product, per shop, per week
- Plain-language questions instead of reports
AI receptionists & chatbots
Customers can ask whether an item is available anywhere in the business and get a real answer, including which shop has it and whether it can be delivered.
- Availability answered across both shops
- Orders and delivery bookings taken in conversation
- Delivery zone and timing confirmed up front
- Works in darja, French and Arabic
Custom websites
A storefront driven by the same live inventory, so the catalogue can never show something the shops do not have.
- Catalogue driven by live stock
- Delivery-zone check and checkout on the page
- French, Arabic and English from one build
- The assistant available on every page
Modeled results
What the model shows
Modeled over two quarters, holding product range and footfall constant.
These figures are modeled for an illustrative retailer of this size. They are not measured results from a named client.
What this looks like for you
You cannot automate answers about stock until the stock itself is in a system. That is usually the first build, and it pays for itself in reordering decisions long before the AI touches a customer.