Learners served
1.2M+
Tutoring agents, assessment platforms, and school-operations AI built around the Arabic learner — aligned with the Ministry of Education's digital agenda and the Human Capability Development Programme. Models that understand the curriculum, the dialect, and the classroom — not Boston-trained chatbots with an Arabic translation layer.
Six pressures we hear from ministry directors, university CIOs, and operators on every first call.
Saudi learners deserve a tutor that thinks in Arabic, understands the local curriculum, and code-switches naturally. Most foreign EdTech models miss the script normalization, the dialect, and the cultural register.
Saudi teachers report 12+ hours per week on planning, marking, and reporting. Lesson generation, structured marking, and parent communications are the highest-ROI AI investment a school can make.
The MoE curriculum is specific, dynamic, and not the same as international frameworks. Anything generated for a Saudi classroom has to be aligned to the actual standard — not a generic adaptation.
Generative AI changed assessment overnight. Schools and universities need authentic-assessment tools, plagiarism-aware grading, and a clear position on what learners can and cannot use AI for — at the policy and the platform level.
Every learner under 18 is a special category under PDPL. Models must be guard-railed for age, language, and content; data must stay in-Kingdom; parental consent must be auditable.
The Human Capability Development Programme defines national skills targets. Tutoring, assessment, and analytics tools have to map to those outcomes — not generic global proficiency frameworks.
Six disciplines, sector-tuned around the Arabic learner, the MoE curriculum, and the school operating model.
A unified data layer across SIS, LMS, and assessment systems. The substrate for personalization, early-warning, and outcome analytics across the school.
Arabic-first, curriculum-aligned tutoring agents — Socratic, not answer-dispensing. Guard-railed for age, language, and the MoE syllabus.
Lesson generation, structured marking, parent communications, and IEP drafting — bilingual, MoE-aligned, and inside the teacher's existing workflow.
In-Kingdom landing zones for school and university workloads, with the integration patterns the MoE / ETEC / TVTC estates actually need.
PDPL-aligned controls for minors, content moderation aligned to MoE policy, and a SOC posture for school networks under increasing threat.
Early-warning, proficiency tracking, and HCDP-mapped outcome analytics — feeding the school leader, not just the dashboard wall.
Two engagements that anchor our EdTech practice. Names redacted under MNDA — the principals and ministry teams know the work.
An MoE-aligned tutoring agent across maths, Arabic, and science for grades 4–9. Socratic by design — the agent does not hand the answer back, it walks the learner toward it. Trained on a Saudi curriculum corpus we built ourselves; guard-railed for age and content.
A bilingual teacher copilot for lesson planning, structured marking, and parent communications — built into the teacher's existing LMS. Outputs are draft-first, teacher-approved; nothing goes home to a parent without a teacher's name on it.
Authorities, school groups, and the integration partners we work with at the institution and ministry level.
Yes — and we say so explicitly. The base model is fine-tuned on a Saudi curriculum corpus we built ourselves, with explicit MoE alignment. We do not wrap a foreign model in an Arabic translation layer.
Age-aware guard-rails on every interaction, content moderation aligned to MoE policy, parental-consent auditing, and PDPL-aligned data handling. Every learner under 18 is treated as a special category, by design.
In-Kingdom, at every step. Sovereign cloud or on-prem depending on the institution's posture. Cross-border learner-data movement is not part of our delivery model.
No. We sit alongside Madrasati, Noor, or whichever LMS / SIS the institution operates. We add the data fabric, the AI workloads, and the teacher-side surfaces — we do not displace the system of record.
Authentic assessment first. Where AI is part of the learner's toolkit, the assessment design has to reflect it; where it is not allowed, the platform enforces it. Policy is co-designed with the institution.
The teacher-copilot suite is the fastest path — most institutions are in pilot inside one quarter. The tutoring agent is a multi-quarter investment because the curriculum-alignment work is real.
Sixty-minute working session with our EdTech lead and a curriculum specialist. Bring the workflow that hurts — tutoring, teacher load, assessment integrity, outcomes — and we'll come back with a one-page roadmap.