Project description
Vocational qualifications go stale because the machinery that maintains them is slow, manual, and expensive. This project puts AI into that machinery: the standards, frameworks, and quality assurance processes that sit upstream of every course and determine whether training systems actually work.
Working with an active national qualifications system, the project has built a suite of AI-powered tools for industry skill bodies, qualifications agencies, and quality assurance teams. The tools span the full pre-delivery pipeline: converting legacy occupational standards into contemporary competency frameworks, generating aligned assessment instruments, and checking qualification architecture for consistency gaps that manual review routinely misses.
The tools also triangulate across multiple data sources: skills frameworks from other jurisdictions, national skills surveys, and international occupational standards, enabling cross-jurisdictional benchmarking and alignment.
The same logic applies to assessment contextualization: a single base standard can be rendered into sector-specific variants across construction, health, and manufacturing in minutes. A human team would take weeks. The underlying problem is familiar: standards and assessment work is slow, expensive, and technically demanding, which means qualifications fall behind industry needs and stay there. AI changes that equation. Built and tested within a live national qualifications system, and grounded in New Zealand’s largest research project on AI-generated vocational assessment, the tools are designed for transfer. Any jurisdiction running competency-based TVET frameworks and under pressure to modernize them is a potential context.
Who is this for
This project is most relevant to you if:
- You work in a ministry, qualifications body, or sector authority responsible for occupational standards or competency frameworks
- You oversee assessment writing and quality or training provider compliance, and learners are already using AI in ways your current processes weren’t designed to detect or handle
- You work for a development agency, multilateral organisation, or technical assistance programme scoping practical, tested AI applications for TVET system modernisation
What we’re looking for
We’re interested in conversations with people working on the same problems in different contexts: a qualifications body exploring AI-assisted standards work, a development programme looking for tools that have already been built and tested, or a system ready to pilot the approach in a new jurisdiction. We are actively seeking implementation partners and co-development opportunities, particularly across Pacific, Southeast Asia, Sub-Saharan Africa, and MENA contexts.
What the evidence shows
These tools were built and tested inside a live national qualifications system, and are grounded in New Zealand’s largest research programme on AI-generated vocational assessment. Of 12,000 standards analysed, 52.5% of Level 5 outcomes were found to be under-pitched.