How Athir Scores
Executive Behaviour
Athir is a standardised, evidence-linked behavioural assessment. This page summarises how the assessment works, what its evidence base is, and what it does not claim. The full detail lives in the Validation Study and the Bias & Fairness Audit.
Scoring Methodology
The assessment
An Athir assessment is an eight-stage behavioural simulation. At each stage the candidate receives staged business communications and interacts with simulated stakeholders: replying to messages, making decisions, delegating, allocating budgets and setting priorities under time pressure. The platform records the full interaction stream, including every response, its timing, and the scenario that prompted it.
The eight competencies
Each recorded interaction is mapped to one of eight competency pillars, each with a documented definition and behavioural indicators used in scoring:
- Strategic Thinking: Vision, long-term thinking, competitive positioning and enterprise strategy.
- Decision Quality: Analytical rigour, judgement under uncertainty and evidence-based decision-making.
- Leadership Influence: Leading people, inspiring performance and building high-performing teams.
- Execution Discipline: Accountability, delegation, operational excellence and consistent delivery.
- Risk Judgement: Ethical decision-making, governance, risk assessment and moral courage.
- Communication Clarity: Executive communication, stakeholder influence and organisational alignment.
- Financial Acumen: Commercial judgement, financial literacy and capital allocation.
- Enterprise Orientation: Customer focus, market awareness and enterprise-wide thinking.
From behaviour to score
The Behavioural Intelligence Engine evaluates recorded behaviour against documented scoring rubrics for the relevant pillar, producing a pillar score on a 0 to 10 scale. The composite score is the unweighted mean of the eight pillar scores: a documented design decision rather than an empirically derived weighting. Readiness ratings and board recommendations are generated by fixed, documented score rules, and they inform human decision-making rather than predicting appointment outcomes.
The same scoring criteria and rubric versions are applied to every candidate regardless of demographic attributes. Whether score outcomes are equivalent across demographic groups in real populations is an empirical question; testing to date is synthetic and preliminary.
The evidence trail
Every score is traceable through a documented chain: assessment output, to competency score, to behavioural indicator, to scored observation, to candidate interaction, to scenario, to the scoring rubric and model versions used. Athir describes this as a traceable evidence architecture: it makes decisions standardised, evidence-linked and reviewable. It does not make them objective in any absolute sense, and Athir does not claim that it does.
Fairness Testing and Evidence Status
Athir's current fairness evidence is an internal audit of 2,847 synthetic responses across five demographic dimensions: preliminary, pre-deployment algorithmic fairness testing. No statistically significant group differences were detected at α = .05 within that synthetic dataset. This does not analyse real candidate data and does not demonstrate the absence of demographic bias in deployed populations.
Criterion-related validity, reliability coefficients and factor-analytic evidence are not currently published; each is being established through documented evidence packages. Athir publishes only claims its evidence supports, and its public documents state current maturity for every evidence area.
Athir is a decision-support tool. No candidate should be selected or rejected solely on the basis of Athir outputs, and appropriately trained humans retain decision authority.