{"slug":"architecture/RECONCILIATION","title":"Reconciliation","section":"architecture","sectionLabel":"Architecture","status":"unclassified","type":"document","path":"architecture/RECONCILIATION.mdx","bytes":3882,"description":"How execution results are verified against intended outcomes — compare, classify, re-reconcile, close the loop.","updated":null,"source":"integratewise/Integratewise-docs-wiki","content":"\nReconciliation closes the loop between execution and canonical truth. Execution is NOT completion. Reconciliation verifies that what was intended matches what actually happened.\n\n## Core Principle\n\n```text\nINTENDED + EXECUTION RESULT + ACTUAL SOURCE STATE\n        ↓\nRECONCILIATION\n```\n\nIf state is unknown:\n\n```text\nSTOP → RECONCILE → THEN decide whether retry is safe\n```\n\n**Rule:** Never blindly retry consequential operations.\n\n## Reconciliation States\n\n| State | Meaning | Action |\n|-------|---------|--------|\n| **Confirmed** | Execution matched intended outcome | Mark complete, update Spine |\n| **Failed** | Execution did not achieve intended outcome | Investigate, retry or escalate |\n| **Unknown** | Cannot determine actual state | Stop, reconcile, then decide |\n| **Divergent** | Actual state differs from intended | Investigate divergence, decide next |\n| **Partially Complete** | Some steps succeeded, some failed | Assess partial progress, decide next |\n| **Pending Reconciliation** | Awaiting source system confirmation | Wait, then reconcile |\n\n## Reconciliation Flow\n\n```text\nEXECUTION COMPLETE\n      ↓\nCOMPARE INTENDED vs ACTUAL\n      ├── Source system state query\n      ├── Receipt analysis\n      ├── Entity version comparison\n      └── Evidence verification\n      ↓\nCLASSIFY RESULT\n      ├── Confirmed → Update Spine, close work\n      ├── Failed → Investigate, retry or escalate\n      ├── Unknown → STOP, reconcile first\n      ├── Divergent → Investigate divergence\n      └── Partial → Assess and decide\n      ↓\nUPDATE SPINE\n      ├── Entity state updated\n      ├── Timeline event emitted\n      ├── Evidence linked\n      └── Memory updated\n      ↓\nNOTIFY USER\n      ├── Outcome reported\n      ├── Evidence provided\n      └── Next work suggested\n```\n\n## Re-Reconciliation\n\nWhen initial reconciliation is inconclusive:\n\n```text\nRE-RECONCILIATION\n      ↓\nDEEPER SOURCE QUERY\n      ├── Query source system directly\n      ├── Compare entity versions\n      ├── Check execution receipts\n      └── Verify evidence chain\n      ↓\nCLASSIFY AGAIN\n      ↓\nDECIDE: RETRY / ESCALATE / CLOSE\n```\n\n**Rules:**\n- Re-reconciliation is bounded — do not loop infinitely\n- Each re-reconciliation attempt is recorded\n- Maximum re-reconciliation attempts are policy-defined\n- After maximum attempts, escalate to human\n\n## Reconciliation and Memory\n\nReconciliation updates organizational memory:\n\n```text\nRECONCILED OUTCOME\n      ↓\nMEMORY UPDATE\n      ├── What worked\n      ├── What didn't work\n      ├── What was learned\n      └── What to do differently next time\n```\n\nThis memory feeds future context assembly, enabling better work allocation and execution over time.\n\n## Reconciliation and the Product Loop\n\nReconciliation completes the product loop:\n\n```text\nSOURCE → SPINE → BUSINESS CONTEXT → PROJECTION → WORKBENCH\n    ↓                                                    ↓\nTWIN → WORK ALLOCATION → GOVERNANCE → CAPABILITY → EXECUTION\n    ↓                                                    ↓\nSOURCE UPDATE ← RE-INGESTION ← RECONCILIATION ← RECEIPT\n    ↓\nUPDATED SPINE → KNOWLEDGE / CONTINUITY → NEXT WORK\n```\n\n## Related\n\n- [Execution Fabric](/architecture/EXECUTION_FABRIC) — what is being reconciled\n- [Spine Schema](/architecture/SPINE_SCHEMA) — what is updated after reconciliation\n- [Organizational Memory](/architecture/ORG_MEMORY_CANONICAL_MODEL) — what reconciliation teaches\n- [Work Graph](/architecture/WORK_GRAPH) — how reconciliation affects work state\n"}