OPERATING DISCIPLINE
The rules should survive the tools.
Platforms change. AI changes. Teams reorganize. DACS principles define what must remain true when decision evidence is evaluated for a stated use, regardless of the implementation used to get there.
They are not a substitute for controls. They are the operating discipline that keeps controls from becoming a checklist detached from scope, evidence, authority, and consequence.
BEFORE YOU DELEGATE THE MECHANICS
Leadership should be able to answer four questions.
Scope
What decision or use are we actually authorizing?
Evidence
What are we allowing that decision to rely upon?
Consequence
What happens when evidence fails or remains unresolved?
Reconstruction
Can we show later why the use was permitted?
Execution can be delegated. The decision boundary still needs an accountable owner.
THE PRINCIPLES
Eight rules keep the decision path accountable.
The principles are deliberately stable. They describe what must remain true even when the technology, review method, operating model, or degree of automation changes.
PRINCIPLE 01
Scope before scale
Readiness is established for a stated decision scope. A result acceptable for one use does not become a blanket endorsement of the enterprise, platform, model, or data.
PRINCIPLE 02
Discovery is not certification
Patterns, classifications, relationships, and AI outputs may support exploration before they are ready for material use. Finding something is not the same as certifying its use.
PRINCIPLE 03
Connectivity is not accountability
Access, lineage, provenance, joins, APIs, and retrieval show that evidence can be reached. They do not establish that it is appropriate for the stated use.
PRINCIPLE 04
Evidence over implementation
DACS does not require a particular catalog, graph, semantic layer, data model, AI platform, or workflow. The requirement is the testable evidence the implementation provides.
PRINCIPLE 05
Control results stay factual
A failed control remains Fail. An approved exception may constrain what happens next, but it does not rewrite the evidence or convert failure into Pass.
PRINCIPLE 06
Readiness must have consequence
A Gate is not just a meeting or note. Its recorded state must permit, constrain, block, or quarantine use according to approved logic for the stated scope.
PRINCIPLE 07
Authority is not execution
A person, workflow, rule, or AI mechanism may execute a control. Execution does not create authority. Ownership, execution, evidence, and recorded outcome remain distinguishable.
PRINCIPLE 08
Reconstruct, remediate, retest
The path must be reconstructable. When evidence changes, affected controls are retested and the Gate record is updated. Fixing the source alone is not enough.
The principles do not require every control to Pass. They require the organization to know what passed, what failed, what remains bounded, and why.
WHY PRINCIPLES MATTER
Practice changes. Accountability should not disappear with it.
A manual review, workflow, policy engine, AI agent, governance platform, or custom control service may all implement parts of DACS differently.
The implementation may vary. The evidence obligation does not.
Principles keep the framework implementation-neutral without making it consequence-neutral.
DACS is vendor-neutral and implementation-neutral. The same decision-evidence obligations must survive technology changes, organizational boundaries, automation, and different operating models.
That neutrality does not mean every implementation is equivalent. A valid implementation still needs explicit scope, inspectable evidence, accountable authority, recorded consequences, and enough preserved context to reconstruct and reassess the decision.
See how existing governance, data, and AI controls can supply evidence >
THE EXECUTIVE PRESSURE TEST
Three reassuring statements that deserve one more question.
Each statement may be true. The principle is to ask whether it establishes readiness for the decision now being made.
“IT IS IN THE APPROVED SYSTEM.”
Approved for which use?
Location and prior approval can be valuable evidence. Neither automatically establishes readiness for the current scope.
“THE MODEL FOUND THE PATTERN.”
Eligible for material use?
Discovery can accelerate learning. The discovered relationship, classification, or output still needs the evidence required for the stated use.
“THE WORKFLOW APPROVED IT.”
Whose authority did it execute?
Automation can execute approved logic. The accountable authority, evidence, constraints, and recorded outcome still need to remain visible.
A strong answer should identify the scope, evidence, authority, controls, and recorded consequence rather than depend on “because the system said so.”
FROM PRINCIPLE TO PRACTICE
The principles tell you what must remain true. The operating model makes it testable.
DACS turns the principles into declared scope, applicable controls, evidence requirements, Gate logic, readiness states, remediation, reassessment, and a preserved record.
GO DEEPER
Read the doctrine
The DACS Manifesto explains the reasoning and boundaries behind the DACS approach.
Test one bounded path
The DACS First Look shows how these principles behave when evidence is actually evaluated.