Method · Evidence

Evidence before paperwork.

The goal is not to prove that documentation exists. It is to obtain enough evidence to justify a conclusion about the knowledge on which important work depends.

Match the evidence and depth of the review to what is at stake. Greater consequences and uncertainty call for closer examination. We still need to test whether this approach achieves a good balance between effort and benefit in practice.

Minimum Sufficient Evidence

Gather the least burdensome evidence set that is still sufficient for a competent and proportionate conclusion. Higher consequence, uncertainty or dependency can require stronger evidence and more challenge.

Reuse before creation

Potential evidence may already exist in skills matrices, process descriptions, quality records, HR information, training records, decision logs, architecture repositories, system documentation, audit material, continuity plans and operational data.

Reuse is not automatic acceptance. Each artefact must support the exact claim in question and be checked for relevance, currency and reliability.

Reuse authoritative results

When a competent external testing or assessment body has already answered the underlying question and its result is usable, ArtiProof reuses that result rather than running parallel conformity checks. Identify the source, version and test or assessment identifier where available. If the result is not usable, record why before undertaking a substitute test or derived assessment.

The ArtiProof review then asks whether the knowledge is fit to be relied upon in the organisation's intended context. Reusing a test result does not automatically establish that fitness.

Keep the original result and an authorised override separate

Some authoritative testing or assessment systems allow an authorised person to override a result. When reusing such evidence, retain the original system-produced result and the governed override as distinct evidence states. Make clear which result supports the decision; an override must never silently erase the original result.

For each override, retain the source, test, version and result identifier; the actor and their authority; the reason or comment; the date and time or version; and the linked supporting evidence. Both states and their relationship must remain traceable.

An override can change the result relied upon only when its authority and rationale can be reconstructed. If that cannot be established, record the unresolved issue rather than treating the override as an authoritative replacement. Whether the evidence is sufficient for the intended use remains a separate judgement.

Know when to stop or refer

ArtiProof does not replace competent legal, fundamental-rights or specialist technical assessment. When a material question falls within such a specialist remit, identify the boundary, preserve the unresolved question and available evidence, and reuse an authoritative specialist result when one is available. Otherwise, stop, qualify the conclusion or refer the question to the appropriate specialist route. Do not give an unqualified confirmation on a matter outside the review's competence.

These rules refine evidence reuse and the boundaries of review. They add neither a sixth assurance property nor a separate conformity or override engine. Their effects on review quality and workload still require independent testing.

Reality testing

Where documents are weak or consequence is higher, evidence can include explanation, observation, walkthrough, interview, peer challenge, exception handling, handover or recovery exercises.

Stopping rule. Discovery can stop when widening the scope would not reasonably change the consequence profile, evidence sufficiency, assurance conclusion or required action. Material exclusions should remain visible.

ArtiProof is ready for research, but has not yet been broadly validated by independent researchers. The first independent replication is still pending. More about the research.