How AI-Powered Assessment is Transforming Youth Development at d-lab
Using Technology to Scale People, Not Replace Them
d-lab is tackling one of South Africa’s most pressing challenges: in the first quarter of 2025, youth unemployment (ages 15 to 34) reached 46.1%. 58.7% of these unemployed youth have no previous work experience. The innovative nine-month d-lab programme transforms unemployed young people into digitally-skilled, design-thinking professionals ready for the Fourth Industrial Revolution (4IR) workplace. With an impressive 85% post programme placement rate, d-lab has proven that their model works. But success brings its own challenge: how do you scale transformative education without diluting its quality or breaking the budget?
The answer lies not in hiring armies of evaluators, but in using artificial intelligence to amplify human capability. At d-lab, we’re pioneering an approach where technology scales people instead of people scaling the business. This philosophy recognises that the most valuable aspects of education – mentorship, creative guidance, and personal development – are inherently human. What AI can do is handle the repetitive, time-intensive tasks that prevent educators from focusing on what they do best: nurturing human potential.
The Assessment Bottleneck: When Success Creates New Challenges
d-lab’s programme is built around portfolio-based assessment – a powerful pedagogical approach where students demonstrate their learning through real-world projects and evidence collections. Throughout their journey from induction through launch, orbit, landing and internship phases, students compile comprehensive portfolios documenting everything from design thinking sprints to financial planning exercises, Canva creations to certification achievements.
This assessment model perfectly aligns with d-lab’s project-based learning philosophy and meets the rigorous standards required by their accreditation with the Institute of Chartered IT Professionals of South Africa (ICITP SA). However, it also creates a significant operational challenge.
Each student submission requires meticulous evaluation across multiple dimensions: technical formatting, proper hyperlinking, content quality, evidence accessibility, and adherence to specific rubrics. Facilitators must check that documents contain proper formatting, verify that evidence links are functional and lead to the correct submissions, ensure stylistic consistency, and provide detailed feedback for improvement.
For many of d-lab’s students – third-language English speakers with high school diplomas entering a demanding professional development environment – this feedback process is crucial. They need specific, actionable guidance on everything from technical document formatting to content quality. But providing this level of detailed assessment manually is extraordinarily time-intensive.
The AI Advantage: Automating the Mundane to Amplify the Meaningful
d-lab’s solution doesn’t replace human judgment – it enhances it. The AI-powered assessment system handles the repetitive, technical aspects of evaluation while ensuring that facilitators can focus on higher-value mentoring and personalised guidance.
The system works by ingesting Google Doc portfolio submissions and subjecting them to a comprehensive multi-stage analysis. First, it extracts and processes document content while automatically removing sensitive personal information to protect student privacy. Next, it conducts detailed formatting analysis, checking everything from table of contents functionality and hyperlink validity to font consistency, margins, and line spacing.
Most importantly, the system evaluates content quality using carefully developed rubric templates that have been refined specifically for AI assessment. These rubrics enable the language model to provide detailed, constructive feedback that’s tailored to each student’s specific submission. The system even verifies that evidence links are accessible and lead to appropriate content.
The output isn’t just a grade – it’s a comprehensive, aesthetically-formatted report that shows students exactly what they did well, what needs improvement, and how to make those improvements. Meanwhile, facilitators receive dashboard insights that help them track student progress and identify those who need additional support.
Beyond Efficiency: The Deeper Impact of AI-Enhanced Assessment
The benefits of this approach extend far beyond operational efficiency. By automating routine assessment tasks, d-lab can:
Democratise Quality Education: Detailed, consistent feedback becomes available to every student, regardless of class size or facilitator workload. This is particularly crucial for students from backgrounds that may have limited exposure to professional-quality feedback in their previous educational experiences.
Enable Rapid Iteration: Students can receive near-immediate feedback on their submissions, allowing them to improve and resubmit quickly rather than waiting weeks for manual evaluation. This accelerated feedback loop is essential for adult learners who need to see rapid progress to maintain motivation.
Maintain Standards at Scale: As d-lab expands its reach, the AI system ensures that assessment quality and consistency remain high regardless of the number of students or geographic distribution of programmes.
Free Human Capacity for Human Work: Facilitators can redirect their time from document checking to what matters most – mentoring relationships, creative guidance, emotional support, and the kind of nuanced feedback that only human experience can provide.
The Future of Scaled Impact
d-lab’s AI-powered assessment system represents more than a technological upgrade – it’s a new model for how educational institutions can achieve massive scale without losing their human touch. By thoughtfully identifying which tasks can be automated and which require human insight, organisations can stretch their impact exponentially.
This approach is particularly relevant for skills-based education in emerging economies, where the demand for quality training far exceeds traditional institutional capacity. As South Africa and similar contexts grapple with massive youth unemployment, the ability to deliver high-quality, personalised education at scale becomes not just an operational advantage, but a social imperative.
The lesson from d-lab’s innovation is clear: the future of educational impact isn’t about choosing between human connection and technological efficiency. It’s about using AI to amplify human capability, ensuring that every young person receives the detailed guidance they deserve while enabling organisations to reach far more learners than ever before.
In a country where almost 59% of unemployed young people have no work experience, that kind of scaling isn’t just innovative – it’s transformational.
By H. Johannes Backer