System: MWMS
Document Type: Protocol
Authority Level: MCR Source Of Truth
Status: Draft For MCR
Primary Location: MCR
Future Operational Destination: HeadOffice Brain, AI Business Systems Brain, AI Manager, AI Employee Router, Brain Room, Course Absorption System, Content Brain, Offer Brain, Sales Brain, Creative Brain, Conversion Brain, Future AIBS Client Systems
Parent Page: HeadOffice
Owner: Martyn
Developer Boundary: No Development Action Authorized By This Page
Source Of Truth: MCR
Purpose
The purpose of this document is to define the MWMS AI Context Activation And Usage Protocol.
This protocol explains how approved MWMS context libraries are activated, loaded, selected, referenced, and used by AI Employees, skills, manual workflows, business asset builds, and future client-facing AIBS systems.
Creating a context library is not enough.
The library only becomes useful when AI Employees know:
when to use it
which files to read
which files not to read
how to apply the context
when to ask for missing context
when to use a skill
when to use a manual build
when to stop and request human review
when to treat a context file as source truth
when to treat an output as draft only
This protocol exists to prevent context libraries from becoming passive documents that sit unused.
MWMS must ensure that approved context is actively applied to real work.
Without this protocol, MWMS risks:
creating context libraries that are never used
AI Employees starting from memory instead of source context
skills triggering without reading the right files
manual builds using incomplete offer knowledge
AI outputs drifting away from approved voice and methodology
client systems using the wrong context
old language resurfacing
proof being invented
offer promises changing without approval
future AIBS systems becoming inconsistent
The AI Context Activation And Usage Protocol turns a stored context library into operational working context.
Scope
This protocol applies whenever an AI Employee, Brain, workflow, skill, or future client system uses approved context to produce output.
This includes:
HeadOffice Brain
AI Business Systems Brain
AI Manager
AI Employee Router
Brain Room
Course Absorption System
Newsletter Intelligence
Opportunity System
Offer Brain
Content Brain
Creative Brain
Sales Brain
Conversion Brain
Customer Brain
Research Brain
Affiliate Brain
Ads Brain
Product Brain
Strategy Brain
future AIBS client systems
This protocol applies to:
offer context libraries
client context libraries
Brain context libraries
AI Employee context packs
skill-based workflows
manual build workflows
asset creation workflows
content generation
ad creation
sales asset creation
lead magnet creation
webinar creation
landing page creation
email sequence creation
client report creation
This protocol does not authorize technical development, plugin changes, Supabase changes, WordPress changes, automation wiring, or M developer action.
Core Definition
AI Context Activation is the process of selecting and applying the correct approved context files before an AI Employee performs meaningful work.
AI Context Usage is the process of using those files correctly while producing outputs.
Context activation answers:
Which library applies?
Which files are relevant?
Which files are active?
Which files are draft?
Which files are retired?
Which files must be ignored?
Which skill should use the library?
Which output is being created?
Which Brain owns the workflow?
Which human review is required?
Context usage answers:
How should the AI Employee apply the source files?
What must be preserved?
What must not be invented?
What must be checked before output?
What should trigger escalation?
Core Principle
The core principle of this protocol is:
AI Employees must use the right approved context before creating important business outputs.
Memory alone is not enough.
A general prompt is not enough.
A skill is not enough if it does not read the right context.
A tool is not enough if it does not understand the source truth.
The approved context library must become the active working base for AI output.
Context Activation Trigger
Context activation is required before creating or reviewing high-value outputs.
Examples include:
ad scripts
VEO3 scripts
landing pages
sales pages
lead magnets
webinars
email sequences
content packs
offer evaluations
client reports
sales scripts
objection handling
brand messaging
campaign briefs
AI Employee instructions
AIBS client deliverables
Context activation may not be required for low-risk brainstorming, early ideation, or casual discussion.
However, once output becomes business-facing or system-facing, context activation becomes required.
Context Source Priority
MWMS uses the following priority order when activating context.
- Current User Instruction
The current user instruction defines the task.
If the user says wait, hold, continue, create full page output, park, ignore, or stop, that instruction governs the session.
- Approved Context Library
The approved context library defines the offer, voice, buyer, method, objections, proof, and retired language.
- Relevant MWMS Standards
Relevant standards define how the work must be performed.
Examples:
Document Structure Standard
Course Absorption Operating Rule
AI Agent Memory And Context Framework
AI Agent Skill Library Framework
Offer Context Library Standard
Context Library Governance And Folder Map Standard
AI Output Validation Standard
- Current Source Material
Current uploaded material, file contents, screenshots, transcripts, or pasted instructions provide task-specific evidence.
- Stable Memory
Stable memory may guide interpretation, but it must not override current instruction, approved context, or source material.
- Assumptions
Assumptions are weakest.
They must be marked or avoided.
Activation Workflow
MWMS uses the following context activation workflow.
Step 1: Identify The Task
The AI Employee must identify what work is being requested.
Examples:
create page output
generate ad script
build lead magnet
review offer
create email sequence
prepare client report
check copy against voice
audit a skill
Step 2: Identify The Owning Brain
The AI Employee must identify the owning Brain.
Examples:
Offer Brain owns offer structure.
Content Brain owns content usage.
Sales Brain owns sales assets.
Creative Brain owns angle and story systems.
HeadOffice owns governance.
AIBS Brain owns future client system packaging.
Step 3: Identify The Relevant Context Library
The AI Employee must determine which library applies.
Examples:
MWMS internal offer library
client library
specific campaign library
specific Brain library
specific project context
If no approved library exists, the AI Employee must flag the gap.
Step 4: Select Relevant Files
The AI Employee must select only the files needed for the current task.
Relevant files may include:
Right-Fit Client Profile
Offer Profile
Voice Architecture
Differentiation Profile
Objection Library
Methodology Map
Expert Thinking Rules
Customer Language Bank
Proof Library
Brand Visual Style
Retired Language
Compliance Notes
Do not load unnecessary files into simple tasks.
Do not ignore necessary files for important tasks.
Step 5: Check File Status
The AI Employee must check whether files are:
Approved
Draft
Under Review
Stale
Retired
Archived
Only approved files should be used as active context unless the user explicitly asks to work from draft material.
Step 6: Check For Retired Language
If the output involves messaging, copy, content, ads, sales, or public-facing assets, Retired Language must be checked where available.
Step 7: Apply Relevant Skill Or Manual Workflow
The AI Employee must decide whether to use:
a formal skill
a manual build
a one-off prompt
a review workflow
a handoff workflow
Skill selection must follow the AI Skill Builder And Audit Protocol.
Step 8: Produce Output
The AI Employee creates the required output using active context.
The output must preserve approved truth and avoid invented details.
Step 9: Validate Output
The output must be checked against the relevant context and standards.
Step 10: Route Or Handoff
The output must have a destination.
Examples:
MCR
HeadOffice review
Brain Room
Content Brain
Offer Brain
Sales Brain
Parking System
Human review
future AIBS client review
Context Selection Rules
Rule 1: Use Only Relevant Context
AI Employees should not load every available file for every task.
Too much context can create confusion.
Use the files needed for the job.
Rule 2: Approved Context Beats Memory
If a context library file conflicts with memory, the approved context file wins.
Rule 3: Current User Instruction Beats Routine
If the user gives a specific current instruction, follow it unless it violates governance or safety.
Rule 4: Draft Context Requires Warning
If only draft context exists, the AI Employee must state that the output is based on draft context and requires review.
Rule 5: Missing Context Must Be Flagged
If a required context file is missing, the AI Employee must not invent it.
Use:
Missing Context
Requires Human Input
Requires Client Confirmation
Requires Source Material
Rule 6: Retired Language Must Be Enforced
If retired language exists, AI Employees must avoid it.
If retired language reappears in output, the output requires correction.
Rule 7: Proof Must Come From Approved Proof Sources
AI Employees must not invent proof, testimonials, results, statistics, customer quotes, or case studies.
Rule 8: Client Context Must Stay Isolated
AI Employees must not use one client’s context for another client.
Rule 9: Project Context Is Temporary
Project context should support the task but should not override approved library context unless promoted.
Rule 10: Context Must Travel Through Handoff
If work moves from one Brain or Employee to another, relevant context must be included in the handoff.
Usage Modes
MWMS recognizes four approved usage modes.
Mode 1: Direct Context Use
The AI Employee reads approved context files and performs the task directly.
Best for:
simple content creation
copy review
offer clarification
message alignment
short-form outputs
low-risk internal drafts
Mode 2: Skill-Based Use
The AI Employee uses a formal skill that reads the relevant context library.
Best for:
repeated workflows
course absorption
newsletter filtering
offer evaluation
content briefs
VEO3 scripts
client reports
developer handoffs
validation checks
Mode 3: Manual Build Use
The AI Employee follows a longer one-off instruction sequence to build a larger asset.
Best for:
lead magnets
webinars
landing page packs
sales page drafts
full funnel asset sets
campaign packages
client deliverable packs
Manual Build Use is appropriate when the output is too specific, too large, or too one-off to require a formal reusable skill.
Mode 4: Review And Validation Use
The AI Employee uses the context library to check an existing output.
Best for:
voice match review
offer accuracy review
proof review
retired language check
objection coverage check
brand alignment review
context drift review
Mode Selection Rules
Use Direct Context Use when the task is simple and context is clear.
Use Skill-Based Use when the task repeats and a formal skill exists.
Use Manual Build Use when the output is large, structured, or one-off.
Use Review And Validation Use when checking existing work against source truth.
Do not create a skill for every manual build.
Do not use manual build mode when a proven skill already exists.
Do not skip validation for public-facing outputs.
Context File Usage Guide
Right-Fit Client Profile
Use when creating:
content
ads
sales scripts
landing pages
lead magnets
webinars
emails
onboarding
client reports
Do not use when the task is only technical or internal unless buyer context matters.
Offer Profile
Use when creating or checking:
offer copy
sales assets
ads
funnels
landing pages
pricing messages
client-facing explanations
Do not invent offer details not present in the file.
Voice Architecture
Use when creating or reviewing:
copy
emails
scripts
social content
client-facing reports
brand-facing output
Do not over-polish if the voice file supports plain, direct, founder-style language.
Differentiation Profile
Use when creating:
positioning
ad angles
sales pages
comparison content
offer explanation
content strategy
Do not turn differentiation into fake superiority claims.
Objection Library
Use when creating:
sales scripts
landing pages
FAQs
VSL scripts
emails
retargeting content
sales enablement
Do not invent objections without evidence.
Methodology Map
Use when creating:
framework explanations
sales mechanisms
course content
client delivery maps
SOPs
AI Employee workflows
Do not simplify methodology into generic advice.
Expert Thinking Rules
Use when creating:
AI Employee logic
diagnostic workflows
review systems
ad analysis
offer analysis
client advice
decision trees
Do not present judgment rules as universal if they are context-specific.
Customer Language Bank
Use when creating:
headlines
hooks
bullets
emails
VSL scripts
landing pages
social content
objection handling
Do not fabricate customer language.
Proof Library
Use when creating:
sales pages
ads
case studies
emails
landing pages
authority content
client reports
Do not invent or exaggerate proof.
Brand Visual Style
Use when creating:
image prompts
thumbnail briefs
banner briefs
visual campaign direction
landing page visual direction
VEO3 scene direction
Do not override permanent brand rules unless approved.
Retired Language
Use when creating or reviewing any public-facing or client-facing messaging.
Do not use retired phrases unless explicitly reviewing history.
Compliance Notes
Use when creating:
ads
sales pages
affiliate content
health claims
finance claims
testimonials
before/after content
client-facing regulated copy
Do not publish or recommend risky wording without review.
Context Activation Failure Modes
MWMS must prevent:
using memory instead of approved context
using the wrong offer library
using the wrong client library
using draft context as approved context
ignoring retired language
inventing proof
inventing buyer language
ignoring methodology
ignoring expert thinking
creating output for the wrong buyer
overloading the task with irrelevant context
routing output without context handoff
using project notes as permanent truth
using old campaign language after repositioning
allowing skills to trigger without reading context
treating a tool as if it understands the business
Output Validation Checklist
Before finalizing an important output, check:
Did the AI use the correct context library?
Were the relevant files selected?
Was the file status approved?
Was retired language checked?
Was proof taken from approved proof sources?
Was the correct buyer used?
Was the offer promise preserved?
Was the voice consistent?
Were objections handled accurately?
Was methodology represented correctly?
Was expert thinking applied where needed?
Were assumptions marked?
Was anything invented?
Is human review required?
Is the output destination clear?
If any answer is weak, the output must be revised, parked, or escalated.
Manual Build Rules
Manual builds should be used when creating large assets that require multiple context files and a deliberate sequence.
Examples:
lead magnet funnel
webinar outline
landing page package
email launch sequence
client report pack
brand messaging guide
campaign asset suite
Manual builds must include:
task objective
required context files
build sequence
output sections
review checkpoints
validation steps
handoff destination
Manual builds should not become permanent skills unless the workflow repeats and passes the Skill Creation Test.
Skill-Based Usage Rules
When a skill uses a context library, it must define:
which context files it reads
which files are required
which files are optional
what to do if files are missing
what it must not invent
what validation is required
how output is routed
A skill that does not define its context requirements is incomplete.
Human Review Requirements
Human review is required when:
context is draft
context is missing
proof is weak
claims are sensitive
client-facing work is produced
public-facing work is produced
paid ad assets are produced
financial decisions are involved
compliance risk exists
developer instructions are involved
automation is considered
MCR pages are being created or changed
Human review protects MWMS from scaling weak context into operational errors.
Context Update After Use
After a context-powered output is created, MWMS should determine whether any new learning should update the context library.
Update may be needed when:
new customer language appears
new objection appears
old positioning fails
new proof is approved
new offer details are clarified
voice correction is made
retired language is identified
a skill repeatedly misuses a file
a campaign result reveals better language
a client corrects assumptions
Context use should feed context improvement.
Governance Role
HeadOffice owns the MWMS AI Context Activation And Usage Protocol.
HeadOffice is responsible for:
ensuring approved context is used before important outputs
preventing memory from replacing source truth
ensuring AI Employees select relevant context
ensuring missing context is flagged
ensuring output validation occurs
ensuring context travels through handoffs
ensuring client context remains isolated
ensuring draft context is not treated as approved
ensuring context updates happen after important learning
Individual Brains may define their own context usage patterns, but they must align with this protocol.
AI Business Systems Brain governs future client-system application.
Offer Brain governs offer-specific context usage.
Content Brain governs content context usage.
Sales Brain governs sales context usage.
Creative Brain governs creative context usage.
Compliance Brain governs claims and risk context usage.
Relationship To Other MWMS Standards
This protocol supports and must align with:
MWMS Document Structure Standard
MWMS AI Agent Memory And Context Framework
MWMS Offer Context Library Standard
MWMS Context Library Governance And Folder Map Standard
MWMS Client IP Excavation Framework
MWMS AI Agent Skill Library Framework
MWMS AI Skill Builder And Audit Protocol
MWMS AI Output Validation Standard
MWMS Messy Input Normalization Framework
MWMS AI Employee Role Card Standard
MWMS AI Employee Capability Stack Framework
MWMS Brain Routing Rule
MWMS MCR Promotion To Brain Protocol
MWMS Page Naming Standard
MWMS Architecture Registry
Content Brain VOC Grounded AI Copy Framework
Research Brain Voice Of Customer Extraction Framework
Offer Brain Offer Structure Framework
Sales Brain Objection Resolution Framework
Creative Brain Belief Shift Framework
Compliance Brain Claims Risk Framework
AI Business Systems Brain Canon
This protocol defines how context moves from stored library into active AI work.
Drift Protection
This protocol protects MWMS from:
unused context libraries
AI Employees starting from generic memory
skills triggering without reading source files
manual builds using incomplete context
wrong offer context usage
wrong client context usage
draft files treated as approved
project notes treated as permanent truth
retired language resurfacing
invented proof
invented customer language
offer promises changing without approval
context not travelling through handoffs
public-facing output created without validation
future AIBS client systems producing inconsistent work
Any important AI output created without confirmed context activation should be treated as a drift risk.
Architectural Intent
The architectural intent of the MWMS AI Context Activation And Usage Protocol is to turn context libraries into active operating intelligence.
MWMS is not building document storage.
MWMS is building a governed AI business ecosystem where AI Employees know what context to read before acting.
The long-term goal is that every important AI workflow can answer:
Which library applies?
Which files matter?
Are the files approved?
What must be preserved?
What must be avoided?
What proof is available?
What language is retired?
What skill or manual workflow should be used?
What output is required?
What validation is needed?
Where does the result go?
What learning should update the library?
When MWMS can answer these questions consistently, AI outputs become more accurate, more aligned, easier to govern, and safer to scale into future client systems.
Change Log
v1.0 — Initial Draft
Created the MWMS AI Context Activation And Usage Protocol as the operating protocol for selecting, loading, applying, validating, and updating approved context libraries during AI Employee work.
This protocol defines context activation triggers, source priority, activation workflow, selection rules, usage modes, mode selection, file usage guidance, failure modes, output validation checklist, manual build rules, skill-based usage rules, human review requirements, context update after use, governance role, drift protection, and architectural intent.
Change Impact Declaration
Pages Created:
MWMS AI Context Activation And Usage Protocol
Pages Updated:
None
Pages Deprecated:
None
Registries Requiring Update:
MWMS Architecture Registry
HeadOffice Page Registry
AI Business Systems Brain Page Registry
Offer Brain Page Registry
Content Brain Page Registry
Sales Brain Page Registry
Creative Brain Page Registry
Canon Version Update Required:
No
Change Log Entry Required:
Yes
Employee Impact Check
Employees impacted:
HeadOffice Manager Employee
AI Manager
AI Employee Router
Context Library Builder
Client IP Excavator
Offer Strategist Employee
Content Planner Employee
Creative Strategist Employee
Sales Strategist Employee
Conversion Strategist Employee
Research Analyst Employee
AI Business Systems Architect Employee
Required behaviour updates:
AI Employees must activate the correct approved context library before creating important offer, client, campaign, content, sales, or public-facing outputs.
AI Employees must select only relevant context files and avoid overloading simple tasks with unnecessary context.
AI Employees must not treat memory, draft files, project notes, or assumptions as approved context.
AI Employees must check retired language, proof sources, buyer profile, offer promise, voice, methodology, and compliance notes where relevant.
AI Employees must flag missing, stale, draft, or conflicting context before producing high-value outputs.
END MWMS AI CONTEXT ACTIVATION AND USAGE PROTOCOL v1.0