How HumanMetrix Measures What Matters
HumanMetrix combines validated psychometric design with AI-powered scoring to measure behavioural fit and practical AI literacy. This page explains how our assessments are built, scored and kept consistent, fair and relevant to the role and region you are hiring for.
The scoring model
Each response is evaluated using an attribution-based model that maps what a candidate says and does to the underlying competency it demonstrates. Scoring is performed by GPT-5 against a benchmarked golden set of reference answers, so an equivalent answer receives an equivalent score every time, removing the drift and rater subjectivity that affect manual scoring.
Short by design, built for prediction
Rather than long inventories of repetitive items, HumanMetrix uses a situational-judgement framework that concentrates measurement on the behaviours and reasoning most predictive of on-the-job performance. The result is a far shorter experience for the candidate that still produces a rich, defensible signal for the hiring team.
The four assessments
FIT measures core behavioural competencies such as agency, curiosity, resilience and growth mindset. AI Readiness measures practical AI fluency. Prompt Architecture measures advanced prompt-engineering and structured reasoning. The Algorithmic Integrity Quotient (AIQ) measures ethical AI understanding, bias awareness and responsible deployment.
Built for every role, not just tech
The same methodology extends to non-technical and safety-critical fields including engineering, construction, healthcare, operations and frontline services. Competencies and thresholds are calibrated to the demands of the role rather than a one-size-fits-all template.
Tailored to your roles
A role is defined by its competencies, each with an importance weighting and a minimum score, together with a job description. Our AI uses that definition to generate the required skill-score profile and to drive role-fit matching, so assessments can be tailored to specialist positions and to roles not yet in the standard library, informed directly by the job description you provide.
Regional relevance and fairness
HumanMetrix supports regionally localised assessment content, and benchmarking, currency and Purchasing Power Parity pricing adapt by country, so candidates are compared against relevant peer cohorts. Scoring against a fixed golden set means every candidate is held to the same standard, and results are reproducible rather than dependent on who is marking.