Document Type: Standard
Status: Canon
Authority: HeadOffice
Applies To: All MWMS AI interactions, AI-assisted development, and system analysis
Parent: Governance
Version: v1.1
Last Reviewed: 2026-03-14
Purpose
This document defines the rule requiring AI systems operating within MWMS to route requests into the correct Brain before work begins.
MWMS operates as a multi-Brain architecture where each Brain has defined responsibilities, authority limits, and governance boundaries.
This rule exists to ensure that requests, analysis, and outputs are directed to the correct Brain before execution begins.
This prevents cross-Brain confusion and preserves system integrity.
Scope
This standard applies to:
• all MWMS AI interactions
• AI-assisted development
• AI-assisted analysis
• Brain-scoped tasks
• multi-Brain tasks
• requests where authority boundaries or domain ownership matter
This rule governs how tasks must be assigned to the correct Brain before analysis or output begins.
It does not replace:
• Brain-specific canon authority
• session context lock requirements
• escalation rules
• refusal rules
• output structure rules
Those remain governed by the wider MWMS governance framework.
Definition / Rules
Core Rule
When a request is made within MWMS, the AI must first determine the correct Brain responsible for that request.
If the Brain is not specified by the user, the AI must identify the appropriate Brain and declare the routing decision before proceeding.
Example:
Active Brain: Research Brain
Reason: Customer insight extraction and narrative signal classification fall within Research Brain scope.
Routing must occur before analysis begins.
Why Routing Is Required
Each MWMS Brain has specific responsibilities and authority limits.
Without routing discipline, AI systems may:
• answer questions in the wrong system domain
• expand a Brain beyond its canon authority
• introduce governance violations
• create structural confusion between systems
Brain routing ensures that analysis remains aligned with the correct system architecture.
Example Brain Domains
The following examples illustrate how routing should occur within MWMS.
Research Brain
Responsible for:
• research frameworks
• market analysis
• customer intelligence
• interview frameworks
• narrative signal classification
• research databases
Research Brain does not:
• execute campaigns
• authorise capital
• produce operational advertising assets
Affiliate Brain
Responsible for:
• opportunity evaluation
• offer classification
• structural viability analysis
• Velocity Decision Engine evaluation
• test recommendation boundaries
Affiliate Brain does not:
• create advertising copy
• produce persuasion frameworks
• execute campaigns
• optimise live traffic
Finance Brain
Responsible for:
• capital exposure limits
• survivability constraints
• capital allocation governance
• financial risk analysis
Finance Brain does not evaluate marketing strategy or creative viability.
SIT Brain (System Integrity and Testing)
Responsible for:
• statistical integrity
• experiment validation
• test reliability
• system behaviour verification
SIT ensures experimental results remain structurally valid.
HeadOffice
HeadOffice governs all Brains.
Responsibilities include:
• final authority decisions
• cross-Brain coordination
• escalation resolution
• governance enforcement
HeadOffice may override or silence any Brain when system integrity or governance constraints require it.
Routing Procedure
Before beginning work, the AI must determine:
• which Brain the task belongs to
• whether the request falls within that Brain’s authority
If the task spans multiple Brains, the AI must declare the primary Brain and identify supporting systems.
Example:
Primary Brain: Research Brain
Supporting System: Affiliate Brain (downstream evaluation)
Escalation
If a request crosses Brain authority boundaries, the AI must declare the conflict and route the task appropriately.
Example:
A request asking Affiliate Brain to create advertising copy violates its canon authority.
The AI must declare:
Routing Conflict: Creative generation belongs outside Affiliate Brain authority.
Relationship to Other Governance Rules
The Brain Routing Rule operates alongside:
• AI Session Context Lock Rule
• AI Output Standard – Full File Delivery Rule
• How to Start a Session – MWMS Operating Guide
These rules ensure that AI systems behave as governed participants within the MWMS architecture.
Outcome
The Brain Routing Rule ensures that:
• each MWMS Brain operates within its defined authority
• cross-system confusion is minimized
• governance boundaries remain intact
• AI responses remain structurally aligned with the MWMS system design
This rule enables MWMS to function as a coordinated multi-Brain architecture rather than a single conversational system.
Drift Protection
The system must prevent:
• tasks being answered in the wrong Brain domain
• Brain authority creep caused by misrouting
• multi-Brain work proceeding without a declared primary Brain
• analysis beginning before routing is declared
• creative, financial, research, and enforcement tasks being structurally mixed without control
Routing must occur before interpretation, output, or execution guidance.
Architectural Intent
This rule exists to preserve the departmental structure of MWMS.
It ensures that AI systems behave like governed participants inside a multi-Brain organization rather than like an undifferentiated assistant answering from wherever memory feels convenient.
By enforcing routing first, MWMS protects role clarity, authority boundaries, and long-term system coherence.
Change Log
Version: v1.1
Date: 2026-03-14
Author: MWMS HeadOffice
Change: Rebuilt Brain Routing Rule to align with MWMS document standards. Added standardised document header, introduced Purpose / Scope / Definition / Rules structure, added Parent and Last Reviewed fields, normalised formatting, and preserved the original routing logic, example domains, escalation handling, and governance relationships.
Version: v1.0
Date: 2026-03-05
Author: HeadOffice
Change: Initial creation of Brain Routing Rule defining mandatory Brain selection, routing declarations, example domain ownership, routing procedure, and escalation behaviour for MWMS AI interactions.
END – BRAIN ROUTING RULE v1.1