A computed figure and a fabricated figure are indistinguishable in text. Output inspection cannot separate them. A tool-call record establishes that a tool executed; it does not establish where the values entering it originated.
Certanum is being built to remove the generative component from the numerical path.
The architecture is designed so that every operand resolves to an authoritative source, through a declared deterministic computation, into a signed record a third party can re-execute. Provenance is established at the operand, not asserted at the output. The output is not the proof. The execution record is.
An operand that cannot be resolved does not degrade the answer. It withholds it. Refusal is a structural property of the architecture, not a policy applied to a model.
Certanum does not establish that a figure is correct. It establishes how the figure was produced.
The generative component is not certified. It is partitioned out of the numerical path — the pattern these regimes already apply to uncertified components. What remains in the path is conventional deterministic software, developable to the assurance bar the domain requires.
The failure is structural rather than sectoral. It occurs wherever a statistical component supplies an operand to a deterministic calculation whose output is acted upon, recorded, regulated or defended — and wherever an existing regime already requires that figure to be evidenced, traceable, validated or reproducible.
Consequential calculation under model-risk control. Independent validation, record-keeping and declared accuracy.
Derived values, validation evidence, and traceability from a derived figure to its predecessor across regulated workflows.
Software assurance and traceability from requirement to implementation to verification evidence.
Reproducibility, traceability and evidence for AI-enabled decision systems under assurance obligation.
Engineering calculation and operational decision where numerical error carries physical or financial consequence.
The architecture is domain-neutral. The initial commercial focus is not.
Descriptions of regulatory obligations are indicative, are not legal advice, and should be confirmed with qualified counsel. Certanum confers compliance with nothing. It produces evidence an existing assurance, validation or audit process may require.
The measurement is open so the claim can be falsified. The system that satisfies it is not.
AP-1, the Admissibility Protocol, is an open standard for measuring whether a consequential figure produced by an AI system was computed, traceable, reproducible and refusable. Authored by Marcus Rupp, published by ZORRZ Financial Inc. under CC BY 4.0, with a reference instrument under MIT. Both are free, and applicable by any party to any system — including Certanum's own, without the publisher's involvement or consent.
Certanum is the proprietary architecture built to satisfy it.
Certanum does not control the standard and holds no certification, registry or accreditation role.