Purchasable reviews for claims, workflows, evidence paths and conformance.
SnapOS does not sell generic AI governance advice. It audits whether a claim, workflow or decision path is still allowed to support a concrete decision.
Can this claim still support the decision?
Scope: one provider/tool/workflow and 3–5 claims. Deliverables: audit memo, claim coverage table, optional evidence gap register and reopen trigger list.
What exactly is the workflow relying on?
Maps model/API, RAG, tools, human review, decision rights, data boundary, monitoring, escalation and reopen triggers.
Can the decision be reconstructed later?
Checks whether a decision path, evidence chain, authority and temporal validity remain independently reviewable.
Who can see, challenge, validate, override and decide?
Maps model access, evidence access, verification rights, decision rights and review asymmetry without premature motive claims.
Can this output become SnapOS-compatible?
Reviews report templates and outputs against the minimal schema, witness set, labels and reliance boundaries.
Is a third-party artefact compatible?
Determines whether an external output meets minimal output and witness rules. Certification is not included.
Typical result labels include SUPPORTED, SUPPORTED WITH CONDITIONS, UNSUPPORTED TRANSFER, REVIEW REQUIRED, DO NOT RELY, OUT OF SCOPE, INSUFFICIENT EVIDENCE and NON-CONFORMANT OUTPUT.