Evidence before opinion.

ATH3NA favours traceable evidence, stated uncertainty and human accountability over claims of certainty.

Evidence-led analysis

Conclusions follow identified evidence rather than unsupported opinion.

Source freshness

Data freshness and availability are carried into every interpretation.

Versioning and corrections

Approved outputs retain version and correction history.

Conflict review

Supporting and contradicting evidence are reviewed together.

Human approval

Commercial editorial outputs require an accountable human review.

Model validation

Model behaviour is assessed against defined outcomes and limitations.

Behavioural observation

Client observations describe recorded conditions and avoid diagnosis.

Explicit limitations

Proxies, incomplete samples and unavailable sources are labelled plainly.

Public methodology does not expose prompts, private evidence, raw provider payloads, internal formulas or private analytical systems.