Demonstrated, not completed.
Every 
certificate is earned through a real performance task — assessed by a trained human against a published rubric. Four certification layers. One consistent standard: proof of judgment, not attendance.
Every certificate is earned through a real performance task — assessed by a trained human against a published rubric. Four certification layers. One consistent standard: proof of judgment, not attendance.
All learners begin at Layer 1. Each subsequent layer builds on the credential before it, going deeper into your sector, your governance role, or your leadership responsibilities.
Each layer earns a distinct certificate naming the exact program, the assessor, and the assessment type. The design is consistent — the credential earned is not.
Certificates are issued digitally with a verifiable certificate ID. Physical copies available on request.
The capstone brief is issued after Module 5. You submit after completing Module 6. One revision opportunity if you fall short of standard on any criterion. All other layers follow the same principle — a real deliverable assessed by a human expert.
Identify one AI system or AI-generated output that you encountered in your professional context during this course. Produce a written evaluation of 600–900 words that:
Learners must meet or exceed standard on at least 4 of 5 criteria. Each criterion is assessed by a trained human assessor — not automated scoring.
| Criterion | Below Standard | Meets Standard | Exceeds Standard |
|---|---|---|---|
| Describes AI accurately | Vague or incorrect description of what the system does | Clear, jargon-free description of the system’s function | Shows understanding of how the system learns and fails, beyond surface description |
| Identifies failure modes | Generic (“AI can be biased”) without specifics | Two specific mechanisms explained with reasoning | Mechanisms tied to SA context and documented evidence |
| Applies Funda Five | Does not apply the course framework | Applies all five steps with reasoning | Extends the framework to context-specific considerations not covered in course |
| Makes a judgment | No clear position taken | Clear judgment with stated reasoning | Judgment addresses competing considerations and remains defensible |
| Accountability | “The company is responsible” (generic) | Named accountability chain with specific role | Identifies gaps in accountability and suggests what should be in place |
If you believe your assessment was not evaluated correctly against the published rubric, you may request a reassessment review within 14 days of receiving your result. Email ask@afriversal.ai with your original submission, the criterion or criteria you are disputing, and a brief explanation of your reasoning. All dispute requests are reviewed by a second assessor who was not involved in the original assessment. You will receive a written response within 10 business days.
AfriversalAI is pursuing QCTO accreditation for the Layer 1 Core Certification, with MICT SETA as Development Quality Partner. Once accredited, employers can recover 40–60% of training costs through the Skills Development Levy discretionary grant system — making AfriversalAI the most price-competitive assessed AI literacy program in South Africa.
QCTO accreditation status: application in preparation — SDL rebate eligibility is pending accreditation approval. Registered organisations will be notified as soon as accreditation is confirmed. Layers 2–4 accreditation applications will follow Layer 1 approval.
