# ADR: Decision Title

| Field | Value |
| --- | --- |
| ADR ID | |
| Status | proposed / accepted / superseded / deprecated |
| Date | |
| Owners | |
| Related systems | |
| Supersedes | |

## Decision contract

| Dimension | Requirement |
| --- | --- |
| Outcome | |
| Operational decision and accepted outcome | |
| System context and affected users/systems | |
| Constraints | |
| Invariants | |
| Quality, latency, risk, adoption, and full-cost ceiling | |
| Accepted residual risk | |
| Revisit trigger | |

## Architecture views

| View | Link or summary |
| --- | --- |
| System context: users, sources, downstream systems, owner | |
| Component boundaries: state, policy, model/ML, tools, UI, operations | |
| Decision/state transitions and postconditions | |
| Deployment/trust boundaries: identity, tenancy, credentials, egress, isolation | |

## Intelligence selection

| Field | Value |
| --- | --- |
| Intelligence-selection record | |
| Deterministic/optimization/ML baseline | |
| Foundation-model or agent rationale, if selected | |
| Human-review and escalation path | |
| Component versions, authority ceilings, monitors, and retirement triggers | |

## Options

| Option | Quality | Latency | Full cost | Security | Operability | Adoption | Lock-in | Evidence |
| --- | ---: | ---: | ---: | ---: | ---: | ---: | ---: | --- |
| | | | | | | | | |

## Selected option

| Field | Value |
| --- | --- |
| Option | |
| Control IDs | |
| Measured evidence | |
| Rejected alternatives | |
| Migration path | |
| Rollback path | |
| Compatibility boundary | |

## Verification

| Test | Threshold | Fixture | Owner |
| --- | --- | --- | --- |
| | | | |