Loading...
Menu
Loading...
Artificial intelligence for the industry of Fez. Computer-vision quality control, supply chain optimisation, production automation and chatbots for the companies working in the city's industrial zones.
Fez has a dense industrial base. The Ain Nokbi and Ben Souda zones hold hundreds of food processing, textile and mechanical engineering firms. A few kilometres away the medina runs the largest concentration of craft workshops anywhere, tanneries, foundries and ceramic studios that have been in production for centuries.
The question is how to bring AI into processes that are still largely manual without wrecking the craft that makes them worth money. Our answer stays narrow: computer vision for grading leather and ceramics, predictive models for food stock, and automation on the textile lines where the work genuinely is repetitive.
The craftsman keeps the craft. The algorithm takes the waste, the rework and the waiting.
A defect model for Fassi zellige is not a defect model for car parts. We train on your material and your own rejects, not on a generic dataset.
No big bang. We pilot on one line or one workshop, measure what changed, then widen it. If the pilot does not pay, nothing else gets built.
Every project trains the people who will live with it. Operators, technicians and managers learn to run the tool, watch it and correct it.
Our AI solutions in Fez
Cameras and deep learning models that catch defects in tanned leather, zellige tiles, woven textiles and packaged food. Scrap rates typically fall by 30 to 60 percent, and the inspection itself runs about ten times faster than a human pass.
Demand forecasting for seasonal goods (dates, olives, winter textiles), stock levels in the Ain Nokbi warehouses, and raw material planning that keeps you off both the stock-out and the overstock.
Trilingual assistants (Arabic, French, English) for Fez firms fielding international enquiries. Quotes, order status and out-of-hours support handled without adding headcount.
Scheduling that keeps machines busy, cuts changeover time and balances the load across the shop. When a machine goes down or a supplier slips, the plan rewrites itself.
From a tannery older than the country to a modern food plant, the tooling changes with what you make.
FAQ
It works, though not the way it does on an automotive line. On a handmade piece you do not teach the model a reference shape, you teach it to recognise a defect, a crack in a glaze or an irregularity in a tanning batch. The variation then becomes information worth keeping rather than a deviation to correct.
Nobody on your side, and that is a condition we set before taking the project on. A tool that assumes a full-time administrator will not survive in an Ain Nokbi workshop, so we ship interfaces that work from a phone and keep maintenance on our side. Training happens at your actual workstations, with your people.
Considerably lower, because the scope is too. A workshop does not need full production supervision, it needs one or two specific automations, usually order handling and batch tracking. We price that narrow scope after a visit, and we say so when a project does not make economic sense at your scale.
The skill of a Fez tanner or ceramicist cannot be automated, and nobody serious claims otherwise. What can be automated is the waste around the work: batches rejected too late, stock badly anticipated, orders rekeyed three times. The craftsperson stays master of the craft while the algorithm handles what pulls them away from it.
Yes, and it is often the fastest project to pay for itself in the Fez industrial zones, because a European buyer will now refuse a batch whose origin is not documented. We link raw material, processing step and shipped batch, so that a recall request is handled in hours rather than days.
Ask for a free read on what you could automate. We name the quick wins and send a realistic AI roadmap within 48 hours.