How the career resilience assessment works

An occupation provides a starting task library. The person taking the assessment keeps, removes and weights tasks by frequency. Each task has reviewed or provisional factors.

The scoring service combines those factors with explicit context answers; an LLM never selects a factor, changes a weight or calculates a score.

The exposure index uses weighted task factors: language digitality 22%, repeatability 18%, tool execution 25%, autonomy feasibility 15% and verification ease 20%.

It estimates how addressable a selected task is under the current model version. It is not a probability of automation, redundancy or unemployment.

Task resilience combines judgement, human interaction, accountability, context variance, physical presence and regulatory constraint, then adds a bounded work-context component.

The score indicates where a responsible person remains important under the model; it does not claim that employment demand will remain unchanged.

Leverage combines augmentability, tool execution, verification ease and the value of retaining human control, plus bounded readiness and verification answers.

It identifies plausible areas for supervised assistance. A high score is a reason to design a safe test, not evidence that a particular product will work reliably.

Selected tasks contribute according to the documented frequency mapping before the result is normalized to 0–100. Confidence is separate from the three indices.

It reflects occupation match, selected-task coverage, editorial coverage, evidence recency and answer completeness, so a numerically precise result can still show limited evidence confidence.

Every result stores its scoring and data versions so later changes do not silently rewrite an earlier report.

Demo factors remain provisional until qualified scoring and occupation reviewers approve them.

The model does not include local vacancies, salary, employer strategy, personal finances or a forecast of future general AI capability.