Research & Methodology

How we build, and what the evidence actually says.

PositionMeAI's tools are built on established career-psychology research, South African validation work, and our own applied methodology. They are not invented from nothing. Some of that grounding is strong. Some of it is still developing. This page explains which is which, and why, rather than asking you to take it on faith.

None of PositionMeAI's tools are validated psychometric instruments. They are structured, developmental reflections, designed and reviewed by registered Industrial Psychologists, not diagnostic or predictive tests. See Meet the Team for who stands behind them.

Established theory & research.

The frameworks below are published, peer-reviewed, or established professional resources, not written by PositionMeAI. Where we use one, we say so on the product itself.

South African evidence.

Grounded in South African samples, not imported wholesale

Potgieter (2012) provides South African early-career validation for the GSAS-derived dimensions (Career Adaptability, Strategic Communication, Learning Agility, Critical Reasoning & Innovation). Maree (2012) provides South African validation for the CAAS (α = 0.91), underpinning the Career Report's Career Readiness module. SAQA's NQF is itself a direct South African regulatory framework. This sits alongside 23+ years of De Vlamingh & Associates' consulting practice in the South African labour market.

Not every construct on the platform has South African-specific validation yet. See Limitations below for exactly which ones.

Our own methodology.

Some parts of the platform are original PositionMeAI work: built for this platform, not drawn from a published instrument. We label these as such wherever they appear, and again here.

How a result actually gets built.

In plain terms: your answers to structured questions are converted into scores across defined dimensions. Those scores are grouped into bands or tiers using thresholds our team has set. Where you see a headline "archetype" or a suggested "route," that's an interpretive label built on top of your scores, meant to help you make sense of a pattern, not a diagnosis. There is no pass, no fail, and no single verdict. No PositionMeAI output is used for hiring, admission, clinical, or credit decisions, and none of them predict an outcome like a job, a hire, or an offer.

Where AI is actually used.

AI drafts. You decide.

The one AI-assisted feature currently live on the platform is the Legacy Scan's personalised AI toolkit. If you use it, your profession, assessment results, and free-text answers are sent to our AI processor, Anthropic PBC, under a commercial API agreement, to generate your toolkit. Anthropic does not train its models on data sent through that commercial API. The CV Generator's AI features are not currently active; if that changes, this page and our Privacy Policy will be updated before the feature goes live, not after.

Full detail on data handling, retention and your choices is in Privacy Policy § 7, AI-assisted features.

Professional oversight.

Every framework above is selected, reviewed and maintained by our team of Industrial Psychologists and Psychometrists. Each is individually registered with the HPCSA, verifiable on the HPCSA iRegister; the practice itself, De Vlamingh & Associates Consulting CC, is not a separately registered entity. AI assists with drafting and personalisation. It does not make the professional judgement calls behind what gets measured, or how a result is framed.

Where the evidence is still developing.

Disclosing a limitation isn't the same as fixing it. Where we can't yet close a gap, we say so plainly instead of overstating what we know.

Honest, specific, current
  • Career Report: interests and values. RIASEC and work-values matching draw on an established global framework that hasn't yet been validated against a South African sample.
  • FuturePath Scan. Its methodology is adapted from research validated on 18–29-year-olds. Applying it to a school-age audience is a deliberate, disclosed extrapolation, not a separately validated instrument for that age group.
  • Legacy Scan. Grounded in the same research tradition as our other tools, but the platform doesn't yet have its own dedicated literature for a retired or experienced-professional population specifically. That's an area we intend to build out.
  • Tier weightings, band thresholds and archetype pairings, platform-wide. These are engineering decisions calibrated by our team, not independently validated statistical models. We say this on the products themselves, and again here.

Ongoing validation.

This is active, evolving work, not a finished state. As the platform grows, our priorities are: closer South African validation of the Career Report's interest and values matching; population-specific grounding for Legacy Scan; and formal validation of our original scales (AI Readiness, Workforce Mobility). We'll update this page as that work happens, not before.