Health systems across lower-middle-income countries in the Middle East and North Africa share a quiet problem: not enough trained, regularly updated clinicians, especially outside capital cities. Perceptors was built to help close that gap. It is an agentic learning management system in which AI agents help institutions author, localise, teach and assess accredited health-worker courses in Arabic and English, with human faculty approving every step.
The problem: a workforce gap that training can't keep up with
WHO projects a global health workforce shortage of 11.1 million by 2030, increasingly concentrated in the African and Eastern Mediterranean regions. In low-income countries there is one health worker for every 621 people, compared with one for every 64 in high-income countries (WHO Executive Board report EB156).
Continuing education hasn't kept pace. Most of it is in person, in English and expensive, so rural doctors, nurses and pharmacists can go years without updated training in non-communicable disease and in endocrine and metabolic care. Patients in poorer and rural communities pay the price, through delayed diagnosis and chronic disease that is poorly managed.
What Perceptors does differently
Instead of a static video library, Perceptors puts specialised AI agents to work alongside human faculty:
- A course-builder agent turns faculty-supplied guidelines and evidence into structured modules, case studies and assessments.
- A localisation agent adapts content into Arabic and the local clinical context, with faculty review before anything is released.
- A tutor agent gives each learner on-demand, evidence-grounded coaching.
- An assessment and analytics agent personalises practice and shows institutions where competency gaps remain.
A typical use case
A ministry of health or a medical association wants to upskill 5,000 primary-care clinicians. Faculty upload the approved guidelines. Perceptors drafts a certified course in Arabic, and faculty review and approve it. Clinicians then learn on their phones in short sessions with AI tutoring, while the institution tracks completion and competency.
How it compares
- Conventional in-person CME is costly and limited by trainer time.
- Generic e-learning platforms are mostly passive, English-first and non-adaptive.
- General-purpose chatbots aren't grounded in approved curricula or accountable to faculty.
Perceptors cuts course-creation time and the cost of delivering each learner's training, supports Arabic natively, and keeps human faculty in control. The platform already hosts healthcare courses certified by the Geneva College of Longevity Science and is live in the United Arab Emirates.
Our theory of change
We focus on three outcomes: clinicians' knowledge and applied competency scores, course completion and retention, and cost per learner trained.
The causal chain is simple:
- Faculty and institutions can build Arabic, guideline-based courses quickly and cheaply.
- More courses reach more clinicians, including rural ones, through mobile access.
- AI tutoring and adaptive practice give the kind of timely feedback live training rarely can, which raises engagement and completion.
- Learners gain knowledge and apply guideline-based practice.
- Over time, better-trained clinicians improve the quality of care for patients.
Digital health professional education is associated with knowledge gains comparable to traditional learning, and tutoring and retrieval practice are associated with better retention. What changes with Perceptors is access (mobile, Arabic, low cost), feedback (immediate and personalised) and faculty capacity (agents handle drafting and administration). This rests on a few assumptions: learners have connectivity and time, faculty approve the content, and institutions adopt the platform and pay for it or find a subsidy.
The pilot we're planning
Next, we plan to pilot Perceptors with partner institutions in Egypt, Jordan, Morocco, Tunisia and Iraq, adapting content to each setting. Using A/B testing, we will compare agentic delivery with standard e-learning on learning gains, completion and cost per learner, alongside a process evaluation of adoption barriers. The results will define the evidence needed for more rigorous testing later.
If your institution, ministry or association trains health workers in the region and wants to take part, we'd like to hear from you.
