Use this when
leaders need to match decision stakes and reversibility to an appropriate level of AI and human authority.
Advanced practitioner depth
Layer 02 · Decisions · Decision Architecture
The Four Decision Tiers
Executive summary
Classify decisions by autonomy, stakes, reversibility, and required human accountability before AI is introduced. This advanced practitioner guide places that work inside Decision Architecture. It helps leaders turn a broad concern into a specific operating decision without treating the topic as a stand-alone transformation. Use the detailed model below to clarify the current state, make trade-offs visible, and assign ownership for the next move. Apply it when leaders need to match decision stakes and reversibility to an appropriate level of AI and human authority. The practical result is a tier assignment for priority decisions with the required human role and control level. Keep that output connected to adjacent layers so upstream constraints remain visible and downstream execution can show whether the design is working.
Practical output
Leave with a tier assignment for priority decisions with the required human role and control level.
Detailed model
How to apply the four decision tiers
Use the practitioner material below after the executive orientation establishes the job, trigger, and expected output.
The Four Decision Tiers
AI belongs in different places depending on stakes, reversibility, and accountability.
A decision tier is not a technology preference. It is an operating judgment about what the organization is willing to automate, assist, supervise, or reserve for humans.
T1 / Autonomous
AI Decides & Acts
Routine, high-volume, easily reversible decisions where AI operates inside a defined corridor.
No human review: A human owns the corridor, monitoring, thresholds, and rollback rules - not each individual decision.
T2 / Assisted
AI Drafts, Human Approves
AI generates recommendations, drafts, or outputs. A human reviews and commits before action.
Human confirms: The approving human owns the decision and cannot transfer accountability to the AI recommendation.
T3 / Supervised
Human Decides, AI Informs
Stakes are meaningful. The human owns the decision while AI provides analysis, data, and scenarios.
Human leads: The human leads judgment, weighs trade-offs, and owns the rationale behind the commitment.
T4 / Reserved
Human Only. No AI.
Irreversible, high-stakes, politically sensitive, ethical, or dignity-related decisions where AI is explicitly excluded from the decision act.
AI excluded: AI may be excluded entirely or limited to background research that does not determine the decision.
Decision Artifacts
A tier becomes real when it is captured as an operating artifact.
Decision architecture should create reusable records, not just workshop language.
Decision tier register
Named Decider map
Reversibility classification
Pre-commit checklist
Decision rationale log
Communication record
Execution owner map
Decision review cadence
Executive Metrics
Decision architecture should improve speed without losing accountability.
Metric
Decision latency
How long does it take from trigger to accountable commitment?
AI can reduce preparation time, but the operating model still needs a clear moment of commitment.
Metric
Reopen rate
How often are decisions re-litigated after they were supposedly made?
High reopen rate signals weak artifacts, unclear deciders, or poor communication of rationale.
Metric
Tier drift
Where is AI performing actions that exceed the authorized decision tier?
Agentic AI raises the consequence of vague decision rights because actions can compound quickly.
Metric
Decision evidence completeness
Can leaders see the evidence, assumptions, tradeoffs, owner, and review date?
Decision memory is the foundation for learning, auditability, and AI-assisted retrieval.
Choose the next path
Return to the layer or apply this topic to the operating model.
The layer overview restores context. The recommended action turns this practitioner model into the next piece of work.