{"slug":"business/DOCTRINE","title":"DOCTRINE","section":"business","sectionLabel":"Business","status":"unclassified","type":"document","path":"business/DOCTRINE.mdx","bytes":11069,"description":null,"updated":null,"source":"integratewise/Integratewise-docs-wiki","content":"INTEGRATEWISE — COMPLETE PRODUCT DOCTRINE\nBuild Brief for Lovable\nVersion: Final — May 2026\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nPART 1 — THE CATEGORY\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\nTHE COMPANY DESCRIPTION\n\nOne line:\nIntegrateWise is the continuity layer for the AI era.\n\nLong form:\nIntegrateWise is the continuity layer for the AI era —\nconnecting systems through a shared protocol, preserving\norganizational memory in a unified Spine, and governing how\nintelligence reasons and acts across tools, models, and time.\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nPART 2 — THE ROOT CAUSE\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\nModern software was built for transactions.\nOrganizations operate on continuity.\n\nTransactions record what happened.\nContinuity preserves why it happened, what it means,\nand how it should influence every future decision.\n\nAs organizations adopt more tools and more AI,\ncontinuity breaks:\n\nContext fragments.\nMemory disappears.\nKnowledge decays.\nDecisions lose lineage.\nHumans become translators between systems.\n\nIntegrateWise exists to restore continuity.\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nPART 3 — THE 18 PROBLEMS\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\nEvery problem below is a continuity failure.\n\n1.  Human is the integration layer\n    → Operational continuity fails\n\n2.  NxN connection explosion\n    → Protocol continuity fails\n\n3.  AI starts cold every session\n    → Memory continuity fails\n\n4.  No persistent organizational memory\n    → Knowledge continuity fails\n\n5.  Stale data across tools\n    → Data continuity fails\n\n6.  Unstructured data has no intelligence\n    → Semantic continuity fails\n\n7.  Writes are ungoverned\n    → Governance continuity fails\n\n8.  Humans don't know what they don't know\n    → Awareness continuity fails\n\n9.  AI executes without human approval\n    → Decision continuity fails\n\n10. System degrades over time\n    → System continuity fails\n\n11. Context lost when models change\n    → Model continuity fails\n\n12. No single operational surface\n    → Operational continuity fails\n\n13. Continuity breaks when things change\n    → All continuity contracts broken simultaneously\n\n14. Knowledge stays static while world evolves\n    → Learning continuity fails\n\n15. Cold start kills adoption\n    → Value continuity fails\n\n16. Best people carry the most institutional risk\n    → Relationship continuity fails\n\n17. Tool lock-in holds data hostage\n    → Tool continuity fails\n\n18. AI wastes tokens on code instead of reasoning\n    → Intelligence continuity fails\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nPART 4 — THE ARCHITECTURE\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\nThree layers. Each necessary. None sufficient alone.\n\nLAYER 1 — PROTOCOL (MCP)\nPurpose: How everything connects.\nSolves: Discovery, invocation, standardization,\nagent-to-agent communication, ecosystem reach.\nMCP is the language.\nGets stronger as MCP ecosystem adoption increases.\n\nLAYER 2 — MEMORY (Spine)\nPurpose: What everything remembers.\nSolves: Context, continuity, organizational memory,\nentity identity, knowledge persistence.\nSpine is the memory.\nModel-independent. Survives tool changes.\nSurvives provider changes. Survives team changes.\n\nLAYER 3 — INTELLIGENCE (Agents)\nPurpose: What everything understands.\nAgents: Twin · Signal Engine · Triage Bot · Operator · Maintainer\nSolves: Reasoning, governance, monitoring,\nproposals, execution, preservation.\nNo Python. No DataWeave. No generated transformation code.\nLLM tokens are reserved for reasoning only.\nDeterministic work is handled deterministically.\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nPART 5 — THE CONTINUITY MAP\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\nEvery component has one job: preserve a type of continuity.\n\nMCP → continuity of communication\nSpine → continuity of memory\nEntity Graph → continuity of relationships\nLineage → continuity of decisions\nSignal Engine → continuity of awareness\nTwin → continuity of human understanding\nTriage Bot → continuity of governance\nOperator → continuity of execution\nMaintainer → continuity of system health\nKnowledge Evol. → continuity of learning\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nPART 6 — THE THREE ENGINES\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\nTHE FACE: Twin (conversation, presentation, action)\nTHE BRAIN: Three Engines (classification, governance, monitoring)\nTHE BODY: Spine (structure, memory, relationships)\n\nEngine 1 — The Normalizer\nWHEN: On arrival — every piece of data that enters\nWHAT: Classifies · Links · Extracts · Files · Triggers\nDOCTRINE: Semantic Contract (Law 2)\nAll data passes through. No exceptions.\n\nEngine 2 — The Triage Bot\nWHEN: On every write — nothing enters Spine without passing through\nWHAT: Deduplicates · Conflict-checks · Scores · Routes · Records lineage\nROUTING:\nHIGH confidence → auto-approve → write to Spine\nMEDIUM → route to human review\nLOW → reject or ask for clarification\nDOCTRINE: Governance (Law 5) + Lineage (Law 7)\nHuman reviews ~20% of writes. 80% flows automatically.\n\nEngine 3 — The Signal Engine\nWHEN: Continuously — always watching\nWHAT: Monitors every entity · Detects deviation ·\nClassifies urgency · Surfaces to Twin\nDOCTRINE: Pattern Detection + Drift Detection\nNot reactive. Proactive. Always on.\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nPART 7 — THE TWO INGESTION PATHS\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\nPath 1 — MCP Path (speaks MCP natively)\nApp → MCP Gateway → Validation → Resolver → Spine\n\nPath 2 — Connector Path (API wrapper)\nApp → Connector → Loader → Normalizer → Resolver → Spine\n\nBoth paths converge at the Resolver.\nBoth paths lead to one destination: the Spine.\nThe path an app takes does not matter to the rest of the system.\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nPART 8 — THE SPINE (TWO LAYERS)\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\nCLOUDFLARE (external interface — what the world sees)\n├── D1 hot cache, governance, fast queries\n├── KV entity cache, TTL 60-300s\n├── R2 blobs, artifacts, exports\n├── Workers compute, protocol, routing\n└── Queues async processing\n\nSPINE DB (internal canonical store — never exposed)\n├── entities\n├── org_memory\n├── conversational_memory\n├── pgvector embeddings\n└── accessed only through Cloudflare Workers\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nPART 9 — THE SEVEN LAWS\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\nLAW 0 — PRIMACY\nThe human is the authority. The system serves human intent.\nNever replaces human judgment.\n\nLAW 1 — PROTOCOL\nEvery agent connects through the Protocol Gateway.\nNo agent connects directly to an external tool.\n\nLAW 2 — SEMANTIC\nEvery piece of data that enters the Spine passes\nthrough the Normalizer. No raw, unnormalized writes.\n\nLAW 3 — CONTINUITY\nEvery agent reads from the Spine first.\nNo agent reads directly from an external tool\nwhen the Spine has the data.\n\nLAW 4 — CAPABILITY\nEvery agent thinks in capabilities, not tools.\nCreate a ticket. Send a message. Update a record.\nThe tools change. The capabilities stay.\n\nLAW 5 — GOVERNANCE\nAI proposes. Human approves. Operator executes.\nNothing runs without human approval.\n\nLAW 6 — MEMORY\nThe human's context is permanent.\nNot dependent on model, provider, or session.\n\nLAW 7 — LINEAGE\nEvery action, every decision, every execution is traceable.\nAccountability serves authority.\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nPART 10 — THE STRATEGIC PROPERTIES\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n1. Deterministic work is handled deterministically.\n   LLM tokens are reserved for reasoning only.\n\n2. No Python. No DataWeave. No generated transformation code.\n\n3. The Spine owns truth. Tools are interchangeable.\n\n4. Human approval is required before anything executes.\n\n5. The architecture gets stronger as MCP adoption increases.\n   (Competitors are threatened by standardization.\n   IntegrateWise benefits from it.)\n\n6. Knowledge survives tool changes, model changes, team changes.\n\n7. The system compounds — every cycle makes it smarter.\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nPART 11 — THE THESIS\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n\"Without MCP, the Spine is isolated.\nWithout the Spine, MCP is just a better REST API.\nWithout Intelligence, both layers sit inert.\nTogether — this is the network layer for organizational context.\"\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nPART 12 — THE OUTCOME\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\nThe organization becomes independent of:\n\n- Which tools it uses today\n- Which AI provider it chose last year\n- Which employees carry institutional memory\n- Which workflows were built by people who have left\n\nKnowledge survives.\nContext survives.\nDecisions survive.\nRelationships survive.\nThe organization survives change.\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\nEND OF DOCTRINE\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n"}