{"slug":"ai/ai-workbench-execution","title":"07 · AI Workbench Execution","section":"ai","sectionLabel":"AI","status":"unclassified","type":"document","path":"ai/ai-workbench-execution.mdx","bytes":4635,"description":"What happens when AI receives work — Business Context, AI Continuity, Active Context, context assembly, reasoning, model selection, planning, capability, governance, execution, observation, continuation.","updated":null,"source":"integratewise/Integratewise-docs-wiki","content":"\n## What Happens When AI Receives Work\n\nThe AI Workbench is a **first-class working environment over business context**. It is not a chat panel, an API payload, or the Twin itself.\n\n```text\nBUSINESS CONTEXT SPINE\n        │\n        ▼\n  Projection (governed, participant-scoped)\n        │\n        ▼\n  ACTIVE CONTEXT BUNDLE\n  Business Context slice · Work · Evidence · Authority · AI Continuity\n        │\n        ▼\n  ┌─────────────────────────────────────────┐\n  │           AI WORKBENCH                  │\n  │  ┌─────────────────────────────────┐   │\n  │  │  Context assembly               │   │\n  │  │  Reasoning & orchestration      │   │\n  │  │  Model routing (AI Gateway)     │   │\n  │  │  Effort / adaptation            │   │\n  │  │  Capability selection (Fabric)  │   │\n  │  │  Planning                       │   │\n  │  │  Governance check (may I?)      │   │\n  │  │  Hermes execution (if authorized)│   │\n  │  │  Observation & continuation     │   │\n  │  └─────────────────────────────────┘   │\n  │              │                        │\n  │              ▼                        │\n  │      AI CONTINUITY MEMORY              │\n  │      (serves Business Context)         │\n  └─────────────────────────────────────────┘\n        │\n        ▼\n  Shared Work · Evidence · Verify → Spine → Rehydrate\n```\n\n\n**Twin lives inside the AI Workbench.** The Twin is the human-facing AI coworker. The AI Workbench is the operating environment.\n\n\n## Internal Flow (derived OODA)\n\nThe Workbench internal planning loop is **OODA-derived** — observe, orient, plan — but it is not exposed as buttons. It is the internal machinery.\n\n\n  \n    AI Workbench hydrates Active Context Bundle: Business Context projection + Work + AI Continuity relevant to this Work. No unrestricted Spine access.\n  \n  \n    Twin/ADK assembles the precise context needed — entity, history, related Work, decisions, and what the participant is authorized to see.\n  \n  \n    Reasoning is scoped to the Work's objective and Business Context. Model selection routes through AI Gateway; effort adapts to Work complexity.\n  \n  \n    A plan is formed: which capabilities, in what order, with what inputs. Capabilities are resolved via [Capability Fabric](/platform/capability-fabric-view) — provider-neutral, not hard-coded tool calls.\n  \n  \n    Before any external mutation, governance answers: *may it be done?* Identity → authority → policy → approval → authorization.\n  \n  \n    Hermes executes the authorized plan. Outcome + evidence return. Verification and reconciliation determine what updates the Spine. Both workbenches rehydrate.\n  \n\n\n## Key Separations\n\n* **Cognition vs execution.** Model routing, agent runtime, capability resolution, authority, and execution remain separated. No monolithic AI Orchestrator.\n* **Twin vs Workbench.** Twin is the coworker. The Workbench is the environment the Twin (and agents, and ADK) operate in.\n* **AI Continuity vs Business Context.** AI Continuity maintains Twin/agent working continuity while remaining grounded in the authoritative Business Context Spine.\n\n## Where to Go Next\n\n<CardGroup cols={2}>\n  \n    The Twin's identity, relationship to user, authority, and memory.\n  \n  \n    Governed organizational actors with identity, role, authority, and evidence.\n  \n  \n    Where this Workbench sits in the end-to-end Work pipeline.\n  \n</CardGroup>\n"}