Methodology FAQ

CoverProof runs a versioned screening methodology against the verbatim Section 250(3) functional test, SM&CR role taxonomy, confidence tiers, and available source context. The output is a decision-support screen for compliance review. Your compliance team makes the final decision before any declaration is sent.

A methodology version is the recorded slug for the rules, prompt, model path, and source-context assumptions used to produce an output. Evidence packs and verifier pages show the version so later reviewers can tell which methodology generated the record.

Benchmark and calibration results are useful only with sample-size and no-certainty caveats. CoverProof suppresses low-count, stale, or mixed-provenance benchmark outputs instead of stretching them into market-wide or legal conclusions.

The verifier checks that a submitted SHA-256 hash matches a ready evidence pack recorded for the firm. It can show methodology version, source-count context, and hash prefixes, but it does not reveal source rows, recipient identities, or free-text submissions.

No. The methodology is structured compliance software and provenance evidence. It is not legal advice, does not decide a firm-specific legal position, and should sit alongside advice from qualified solicitors where the firm needs legal judgment.

A methodology change needs a version slug, fingerprint, evaluation evidence, drift status, claim-gate status, and immutable audit record. Promotion remains blocked while evaluation, drift, or public-claim requirements are unresolved.