10. What DACS Does and Does Not Claim

10.1 What DACS Defines

A vendor-neutral, implementation-neutral control specification for decision-evidence readiness within a stated decision scope.
Profiles, control criteria, evidence obligations, control results, Gate readiness states, and claims boundaries.
The distinction between discovery, governed review, constrained use, Quarantined evidence, and stated decision use.
Requirements for reconstructable evidence paths, structured Conformance Records, and controlled exception handling.

DACS does not guarantee that a decision is correct, optimal, lawful, risk-free, or immune from challenge.
DACS conformance is specific to a stated decision scope. It does not establish organization-wide conformance, a DACS Capability Level, formal certification, or legal or regulatory compliance status. It also does not certify an enterprise, platform, vendor, model, repository, or product.
DACS does not replace data governance, data quality, metadata management, taxonomy, ontology, lineage, AI governance, risk management, architecture, or stewardship.
DACS does not require one data model, graph, catalog, semantic layer, warehouse, lakehouse, workflow platform, AI platform, or other implementation architecture.
DACS does not make connectivity, availability, provenance alone, prior approval, or technical accessibility equivalent to decision readiness.
Execution by an automated rule, Gate, workflow, or model does not make that mechanism the accountable authority.

DACS requires testable control evidence. The following conditions do not, by themselves, establish readiness for a stated use:

unmanaged inference presented as enterprise truth;
identity, classification, and relationship assumptions hidden inside reports, pipelines, dashboards, semantic layers, model features, or agent prompts;
confidence without evidence;
classification without governance;
association without orientation;
relationship vectors that describe risk but do not govern use;
lineage that cannot explain decision logic;
human approval without inspectable evidence;
AI outputs that drift from governed meaning, approved classifications, permitted relationships, or decision-scope context without detection, review, or consequence; and
governance that cannot be implemented, observed, tested, or evidenced.

DACS permits discovery, analytics, automation, and AI. It prohibits treating untested discovery as evidence ready for material decision use.