AI that ships in weeks. And a team that runs it.
We build AI products and agents with you, then train your people to run them. From Casablanca, in French, Arabic and English.
- AI agents
- RAG
- Automation
- Training
- Claude
- Evaluation
- Local models
- Activations
Every company has an AI deck. Few have an AI that works.
Pilots that never ship. Tools nobody adopts. Advisers who leave a slide deck and an invoice. The gap isn't the technology; it's getting from an idea to something your team uses every day.
Four crafts. One studio.
Experiences spark interest, Academy builds skills, Labs builds the product, Apps scales it. Drag a totem to turn it.
The Khatam
Labs builds: audits, sprints, ready-made accelerators and custom builds. Eight points, one working core.
See solutionsWeeks, not quarters.
- Week 1
A working prototype
Real screens on your data. A clear go or no-go.
- Weeks 2–6
Production
Built with your team in the room, evaluated, deployed.
- Then
We train your team to own it
Handover through the Academy, so it keeps running without us.
The tools we build with.
From frontier models to models hosted on your own servers: we pick the tool for the problem, not the other way round.
- Claude
- Claude Code
- OpenRouter
- Open models
- Local models
- AI agents
- RAG
- Evals
- n8n
- Make
- Next.js
- Vercel
- Python
Built alongside people who deliver.
SkillLens
What the room understood, before and after.
Our training-assessment SaaS. From quiz to proof.
Visit SkillLens- Writing a promptBefore 32%, After 78%
- Checking an answerBefore 41%, After 74%
- Protecting dataBefore 28%, After 81%
- Automating a taskBefore 22%, After 63%
We have the answer.
How do we know if our need is ready for AI?
That is what the first call is for: we look at the process, the decisions, the exceptions and the data. Reading and answering from your documents is RAG; deciding and acting is an agent. If AI isn't the answer, we tell you.
Is our data used to train models?
No. We work with providers that don't keep your data or train their models on it, and, when confidentiality calls for it, with models hosted on your own servers.
What does it cost?
We don't publish prices. After the call, each project is priced in a written proposal, and nothing binds you until both parties have signed it.
Who owns what you build?
Labs: everything we build for you is yours. Apps: the product stays ours; your data and your workspace stay yours. Academy: the content stays ours; you keep lifetime access, with no right to resell it.
What happens after launch?
We train your team to run it, through the Academy. If you prefer, we take care of monitoring, evaluations and iterations on a subscription (Run).