The Control Stack
Reference Surface
This page is part of the canonical Decision Infrastructure reference model.
The architecture of governed consequence
Seven layers. One causal chain — from strategic intent at the top to evidenced outcome at the bottom. Each layer reasons, governs, forms, or tracks. Only one determines execution admissibility.
That one is Decision Infrastructure — the only runtime gate in the stack.
The architecture of governed consequence
The Control Stack is not enterprise reference architecture. It is the lifecycle of how enterprise systems produce, govern, and learn from consequence.
At the top, strategic intent. At the bottom, evidenced outcome. Between them, the chain of governance that determines whether a decision becomes a valid action.
Most enterprise architectures skip one layer in that chain. That single omission is where execution integrity breaks — and where downstream intelligence becomes untrustworthy.
The seven layers
Each layer answers a single question. The questions, in order, trace the causal flow from intent to outcome.
Strategic Alignment
Mandate validation · problem framing · strategic intent
Are we solving the correct problem?
Trust & Governance
Data governance · metrics governance · AI governance
Are governance conditions trustworthy?
Sovereign Reasoning
Safe inference · boundary enforcement
What is the system allowed to reason about?
Decisioning
Rules · policies · constrained model outputs
What decision is being formed?
Decision Systems
Traceability · artifacts · lifecycle management
How is the decision operationalized?
Decision InfrastructureRuntime Layer
Runtime admissibility · execution governance
Is execution still admissible right now?
Consequence Intelligence(Output)
Governed outcomes · evidence · continuous learning
What was learned from governed execution?
Where execution becomes irreversible
Between Layer 5 (Decision Systems) and Layer 6 (Decision Infrastructure) sits the only labeled transition in the stack: the commit boundary.
The question that crosses it is operational, not architectural:
Commit Boundary
“Should execution proceed?”
Before that moment, a decision is reversible. After it, it is not. The commit boundary is where intent becomes consequence — and the only structural point where execution governance can refuse before action becomes real.
Why Decision Infrastructure is the runtime gate
At Layer 6, the system continuously revalidates admissibility at the moment of action. The check is structural and runtime:
Revalidates
state · policy · authority · constraints · evidence
Every commit resolves into one of three states — deterministically, before irreversible transition occurs.
Every layer above L6 reasons, governs, forms, or tracks. None of them refuse. The runtime gate exists because admissibility at approval is not the same as admissibility at execution — and the system that decides them must be different from the system that produced them.
Why learning integrity depends on admissible execution
Layer 7 — Consequence Intelligence — is not a system you deploy. It is the output of the Control Stack. The quality of L7 depends entirely on what is allowed to reach it.
When the runtime gate is bypassed, L5 outputs feed L7 directly. The optimization loop receives signal from execution that may never have been admissible: stale state, drifted policy, expired authority, incomplete evidence — all entering the training and feedback loop as if it were governed truth.
The learning loop becomes corrupted at its source.
When L6 governs, only admissible execution proceeds. Layer 7 receives signal only from executions that were valid at the moment they occurred. The optimization loop becomes structurally trustworthy.
Consequence Intelligence is only trustworthy
if execution was admissible.
That is why Decision Infrastructure is not optional. It is the structural primitive that makes everything above it trustworthy — and everything downstream of it valid to learn from.
How Control Tower operationalizes visibility
The Control Stack governs execution. Control Tower makes that governance visible.
Operators see governed execution live — every admit, every deny, every hold — with the reason attached. Not as activity logs. Not as dashboards reconstructed after the fact. As the real-time state of L6 itself.
That is the difference between observability and visibility. Observability watches the system. The Control Tower shows the runtime layer doing what it does.
Every layer reasons, governs, forms, or tracks.
Only Decision Infrastructure determines execution admissibility.
Relationship Reading Tree
Relationship to Other Concepts
Decision Infrastructure is part of a connected ontology. Use this relationship tree to understand where this concept fits.
- System of Intelligence
- Decision Infrastructure
- Decision-to-Execution Gap
- Commit Boundary
- Execution Governance
- Runtime Admissibility
- Governed Execution
- Evidence at Execution
- Operational Legitimacy (Result)
- Consequence Intelligence (Output)
Reference Surfaces
Architecture Surfaces
Architectural reference indexes
Architecture anchors that explain how Decision Infrastructure operates — distinct from the canonical anchor pages above and the ontology spine.
Reference Surfaces
Reference Surfaces
Understanding a category requires more than comparisons. These reference surfaces explain the core concepts, architecture, vocabulary, and placement of Decision Infrastructure within the enterprise stack.
Definition
What Is Decision Infrastructure?
The canonical introduction to the category. Defines Decision Infrastructure, execution governance, runtime admissibility, and governed execution.
- Category definition
- Execution governance
- Runtime admissibility
- Governed execution
Placement
Where Decision Infrastructure Fits
Where Decision Infrastructure sits between Decision Systems and Consequence Intelligence in the enterprise stack.
- L4 Decisioning
- L5 Decision Systems
- L6 Decision Infrastructure
- L7 Consequence Intelligence
Architecture
Decision Infrastructure Architecture
The architecture that enables execution governance — how Decision Infrastructure operates across enterprise systems.
- Commit boundaries
- Runtime validation
- Execution control
- Evidence generation
Vocabulary
Decision Infrastructure Glossary
The canonical vocabulary of the category — the lexicon analysts can quote precisely.
- Runtime admissibility
- Commit boundary
- Execution governance
- Governed execution
- Evidence at action
Related Concepts
Architectural primitives within the Control Stack
The architectural primitives that compose Decision Infrastructure — each governs one facet of how execution remains admissible — plus the operating model that runs them.
Commit Boundary
The structural point where intent crosses into consequence — the runtime gate at L6.
Execution Governance
The discipline of controlling execution at the moment decisions become consequences.
Runtime Admissibility
The property that an approved decision remains permitted at the moment it acts.
Governed Execution
Execution that occurs only when policy, authority, conditions, and evidence remain valid at the act.
Evidence at Execution
Evidence captured at the moment of action — not reconstructed afterward.
System of Intelligence
The operating model that runs the Control Stack — how Decision Infrastructure operates across enterprise state.
Governance Ontology
The semantic layer admissibility evaluations operate on — the shared vocabulary that makes L6 verdicts comparable across enterprise systems.
Related Comparisons
Related Comparisons
Use these comparisons to understand how Decision Infrastructure differs from adjacent categories, systems, and governance models.
Decision Infrastructure vs Decision Intelligence
The category vs its output cousin — what produces decisions vs what governs them at execution.
Decision Infrastructure vs Decision Governance
Governance defines policy. Infrastructure operationalizes it at execution.
Decision Infrastructure vs Decision Control Plane
A control plane routes and coordinates actions; Decision Infrastructure governs whether each action should still happen at all.
Decision Infrastructure vs Decision Execution Engine
An execution engine runs the action; Decision Infrastructure governs whether execution may proceed.
Decision Infrastructure vs Runtime Governance
Runtime governance is a capability; Decision Infrastructure is the category that contains it.
Decision Infrastructure vs Decision Systems
Workflow-and-approvals systems exit before execution; Decision Infrastructure governs the act itself.
Decision Infrastructure vs AI Governance
AI Governance defines what should be allowed. Decision Infrastructure governs whether those permissions remain valid at execution.
AI Governance vs Decision Systems
Why model and process governance frameworks don't close the gap between approval and consequence.
Decision Infrastructure vs Digital Twin
Simulating reality vs governing what is allowed to happen in reality.
Sovereign Reasoning vs Decision Systems
Reasoning under jurisdictional and policy constraints vs the workflow systems that operationalize decisions.
Decision Infrastructure vs Agentic AI
Agents act autonomously; Decision Infrastructure governs whether each autonomous action is admissible at execution.
Decision Infrastructure vs MLOps
MLOps keeps the model healthy; Decision Infrastructure governs whether the decision it informs is admissible at execution.
Decision Infrastructure vs GRC
GRC documents and reviews controls; Decision Infrastructure enforces them on each action at execution.
Decision Infrastructure vs iPaaS
iPaaS connects systems and moves data; Decision Infrastructure governs whether the action between them should execute.
Decision Infrastructure vs Observability
Observability explains execution; Decision Infrastructure governs whether it should occur at all.
Decision Infrastructure vs Knowledge Graphs
Knowledge graphs map what is connected; Decision Infrastructure governs whether an action across those connections is admissible.
Decision Infrastructure vs Sovereign Reasoning
Sovereign Reasoning bounds how AI reasons; Decision Infrastructure governs whether the resulting action is admissible at execution.
Decision Infrastructure and Palantir
Palantir integrates data and drives action; Decision Infrastructure governs whether each action is admissible at execution — across any platform.
Decision Infrastructure and ServiceNow
ServiceNow runs and automates the workflow; Decision Infrastructure governs whether each action it fires is admissible at execution.
Decision Infrastructure and Pega
Pega manages decision workflows; Decision Infrastructure governs whether execution remains legitimate at the act.
Decision Infrastructure and Appian
Appian automates process execution; Decision Infrastructure governs consequence authorization at the commit boundary.
Decision Infrastructure and FICO
FICO optimizes decision quality; Decision Infrastructure governs whether a scored decision is still admissible at execution.
Decision Infrastructure vs Middleware
Middleware passes messages between systems; Decision Infrastructure governs whether the action a message triggers should execute.
Decision Infrastructure vs BPM
BPM orchestrates the process and moves work to the action; Decision Infrastructure governs whether that action should commit.
Decision Infrastructure vs Workflow Automation
Workflow automation runs the sequence; Decision Infrastructure governs whether each action in it should commit.
Decision Infrastructure and Salesforce
Salesforce runs the customer workflow; Decision Infrastructure governs whether each action it fires remains legitimate at the act.
Decision Infrastructure and Celonis
Celonis reveals how processes run and drives action; Decision Infrastructure governs whether that action is admissible at execution.
Decision Infrastructure and Icertis
Icertis manages contracts and obligations; Decision Infrastructure governs whether an action taken under them is admissible at execution.
Decision Infrastructure and Encompass
Encompass runs the loan workflow; Decision Infrastructure governs whether each consequential loan action is admissible at execution.
Decision Infrastructure and Empower
Empower runs loan origination; Decision Infrastructure governs whether each consequential loan action is admissible at execution.
Decision Infrastructure and Harvey
Harvey generates legal reasoning and drafts; Decision Infrastructure governs whether the actions taken from that reasoning are admissible at execution.
Decision Infrastructure and iManage
iManage manages legal knowledge; Decision Infrastructure governs the consequential actions taken using that information at execution.
Decision Infrastructure and Intapp
Intapp coordinates legal intake, conflicts, and approvals; Decision Infrastructure governs whether execution remains admissible at the act.
Decision Infrastructure and Relativity
Relativity surfaces and reviews evidence; Decision Infrastructure governs the consequential actions taken because of it at execution.
Decision Infrastructure and Reveal
Reveal surfaces evidence with AI-assisted review; Decision Infrastructure governs the consequential execution based on it.
Decision Infrastructure and Aderant
Aderant runs the business of law; Decision Infrastructure governs whether the consequential actions those operations drive are admissible at execution.
Decision Infrastructure and NetDocuments
NetDocuments manages legal documents and knowledge; Decision Infrastructure governs the consequential actions taken using that information.
Decision Infrastructure and Contract Lifecycle Management
Contract lifecycle platforms manage the contract; Decision Infrastructure governs whether actions taken under it remain admissible at execution.
Decision Infrastructure and Litera
Litera drafts, compares, and perfects legal documents; Decision Infrastructure governs whether the actions taken from those documents are admissible at execution.
Related Reading
Long-form explorations of the Control Stack
Platform & Vision