Intelligence Domain Events¶
Why This Exists¶
This document defines domain events published by the Intelligence Domain.
Events allow Intelligence to interact with Organization, Event Mesh, Procurement Execution Engine, audit, analytics, and future domains without taking ownership of their data.
Owner¶
The owner is the Chief Product Officer and Enterprise Architect.
Business Value¶
Intelligence events make AI behavior observable, governable, explainable, and auditable.
Event Metadata¶
Every Intelligence event should include:
- Event ID.
- Event name.
- OrganizationId where customer-bound.
- Session ID.
- Agent ID or Digital Professional ID.
- Correlation ID.
- Causation ID.
- Timestamp.
- Confidence score where applicable.
- Guardrail result where applicable.
- Audit reference.
Event Catalogue¶
| Event | Trigger | Payload | Consumers | Business value |
|---|---|---|---|---|
| IntelligenceSessionStarted | Reasoning work begins. | SessionId, OrganizationId, objective. | Audit, Analytics. | Makes AI work observable. |
| OrganizationContextLoaded | Organization context is loaded for reasoning. | OrganizationId, context references. | Audit, AI Memory. | Shows source context used. |
| AgentSelected | Intelligence Director selects an agent. | AgentId, DigitalProfessionalId, reason. | Audit, Analytics. | Explains orchestration. |
| ReasoningCompleted | Reasoning session produces output. | SessionId, output summary, confidence. | AI Memory, UX, Analytics. | Records intelligence result. |
| ConfidenceAssessed | Confidence score is assigned. | Score, method, threshold. | UX, Audit, Guardrails. | Supports trust and review. |
| HumanApprovalRequired | Guardrail requires approval. | RequestId, approver role, reason. | Notification, UX, Execution Engine. | Preserves human leadership. |
| HumanApprovalGranted | Human approves action. | Approval ID, approver, conditions. | Execution Engine, Audit. | Allows governed execution. |
| HumanApprovalRejected | Human rejects action. | Approval ID, reason. | AI Memory, Analytics. | Improves learning and trust. |
| ExecutionRequested | Intelligence requests workflow execution. | RequestId, action, OrganizationId. | Procurement Execution Engine. | Connects reasoning to action. |
| ExecutionBlocked | Guardrail blocks execution. | RequestId, policy, reason. | Audit, UX. | Prevents unsafe automation. |
| AIMemoryUpdated | Intelligence memory is updated. | MemoryId, source, scope. | AI Reasoning, Analytics. | Supports learning. |
| ExplainabilityRecorded | Explanation is captured. | ExplanationId, evidence references. | UX, Audit. | Improves transparency. |
| IntelligenceAuditLogged | Audit record is written. | AuditId, action, correlation ID. | Audit, Compliance, Admin. | Supports governance. |
Event Flow¶
flowchart LR
EventMesh[Event Mesh]
Intelligence[Intelligence Domain]
Reasoning[Reasoning Session]
Guardrails[Guardrails]
Execution[Procurement Execution Engine]
Memory[AI Memory]
Audit[Audit]
EventMesh --> Intelligence
Intelligence --> Reasoning
Reasoning --> Guardrails
Guardrails --> Execution
Reasoning --> Memory
Reasoning --> Audit
Execution --> EventMesh