Reconciliation

Reconciliation closes the loop between execution and canonical truth. Execution is NOT completion. Reconciliation verifies that what was intended matches what actually happened.

Core Principle

INTENDED + EXECUTION RESULT + ACTUAL SOURCE STATE
        ↓
RECONCILIATION

If state is unknown:

STOP → RECONCILE → THEN decide whether retry is safe

Rule: Never blindly retry consequential operations.

Reconciliation States

State	Meaning	Action
Confirmed	Execution matched intended outcome	Mark complete, update Spine
Failed	Execution did not achieve intended outcome	Investigate, retry or escalate
Unknown	Cannot determine actual state	Stop, reconcile, then decide
Divergent	Actual state differs from intended	Investigate divergence, decide next
Partially Complete	Some steps succeeded, some failed	Assess partial progress, decide next
Pending Reconciliation	Awaiting source system confirmation	Wait, then reconcile

Reconciliation Flow

EXECUTION COMPLETE
      ↓
COMPARE INTENDED vs ACTUAL
      ├── Source system state query
      ├── Receipt analysis
      ├── Entity version comparison
      └── Evidence verification
      ↓
CLASSIFY RESULT
      ├── Confirmed → Update Spine, close work
      ├── Failed → Investigate, retry or escalate
      ├── Unknown → STOP, reconcile first
      ├── Divergent → Investigate divergence
      └── Partial → Assess and decide
      ↓
UPDATE SPINE
      ├── Entity state updated
      ├── Timeline event emitted
      ├── Evidence linked
      └── Memory updated
      ↓
NOTIFY USER
      ├── Outcome reported
      ├── Evidence provided
      └── Next work suggested

Re-Reconciliation

When initial reconciliation is inconclusive:

RE-RECONCILIATION
      ↓
DEEPER SOURCE QUERY
      ├── Query source system directly
      ├── Compare entity versions
      ├── Check execution receipts
      └── Verify evidence chain
      ↓
CLASSIFY AGAIN
      ↓
DECIDE: RETRY / ESCALATE / CLOSE

Rules:

• Re-reconciliation is bounded — do not loop infinitely
• Each re-reconciliation attempt is recorded
• Maximum re-reconciliation attempts are policy-defined
• After maximum attempts, escalate to human

Reconciliation and Memory

Reconciliation updates organizational memory:

RECONCILED OUTCOME
      ↓
MEMORY UPDATE
      ├── What worked
      ├── What didn't work
      ├── What was learned
      └── What to do differently next time

This memory feeds future context assembly, enabling better work allocation and execution over time.

Reconciliation and the Product Loop

Reconciliation completes the product loop:

SOURCE → SPINE → BUSINESS CONTEXT → PROJECTION → WORKBENCH
    ↓                                                    ↓
TWIN → WORK ALLOCATION → GOVERNANCE → CAPABILITY → EXECUTION
    ↓                                                    ↓
SOURCE UPDATE ← RE-INGESTION ← RECONCILIATION ← RECEIPT
    ↓
UPDATED SPINE → KNOWLEDGE / CONTINUITY → NEXT WORK

Related

• Execution Fabric — what is being reconciled
• Spine Schema — what is updated after reconciliation
• Organizational Memory — what reconciliation teaches
• Work Graph — how reconciliation affects work state