Methodology

Competency Labs does not build assessments from opinion. We start with published competency frameworks, regulatory expectations, and employer-defined job tasks. Scenarios are then written by subject-matter experts and reviewed against real-world failure modes that hiring managers report.

• FrameIdentify the role, tasks, and decisions a professional must make on the job.
• ScenarioWrite realistic, open-ended situations that require judgment, prioritization, and communication.
• RubricDefine observable criteria for competency and entrustability, aligned to the framework.
• Pre-scoreAI scores against the rubric to produce a consistent first-pass signal.
• Human reviewTrained reviewers validate or adjust the CompetencyScore™ before it becomes verified evidence.
• PublishVerified evidence is added to the Competency Passport with full audit metadata.

Every response is scored on two independent layers: competency (does the candidate demonstrate the skill) and entrustability (how much supervision would be required). This pairing gives employers a richer signal than a single score can provide.

We track item-level performance, reviewer agreement, and score stability across attempts. Rubrics and scoring models are versioned; once a Passport is issued, its underlying evidence is frozen so the credential remains auditable over time.

The research team does not report to the commercial team. Assessment quality is reviewed independently of growth targets. No score is issued from AI alone, and development-mode results are never mixed with verified evidence.