System: MWMS
Document Type: Operating Framework
Authority Level: MCR Source Of Truth
Status: Active
Version: v1.2
Primary Location: MCR
Future Operational Destination: Sales Brain, AIBS Brain, PPL Brain, Affiliate Brain, Automation Brain, Data Brain, Compliance Brain, Risk Brain, HeadOffice Brain
Parent Page: Sales Brain
Owner: Martyn
Developer Boundary: Do Not Touch M’s Active Build Areas Unless Specifically Assigned
Source Of Truth: MCR
Last Reviewed: 2026-06-21
Source / Origin: MWMS AI Assisted Outreach And Sales Follow Up Automation Framework v1.1 and AI Automations by Jack material covering prospect research, company intelligence, RAG-assisted communication, personalised outreach, lead enrichment, reply classification, appointment setting, CRM movement, proposal follow-up, voice-note style outreach, and multi-stage sales automation
MWMS Classification: Sales Brain Operating Framework / AI Assisted Outreach Standard / Sales Follow Up Automation Framework / Prospect Intelligence To Communication System / Human Reviewed Outbound Governance Framework
Primary Brain: Sales Brain
Supporting Brains: AIBS Brain, PPL Brain, Affiliate Brain, Research Brain, Data Brain, Automation Brain, Compliance Brain, Risk Brain, Content Brain, Customer Brain, Experimentation Brain, HeadOffice Brain
Related Pages: Sales Brain Canon, AIBS Brain Canon, MWMS Outbound Lead Enrichment And Cold Outreach Governance Framework, MWMS Lead Intake Qualification And Follow Up Automation Framework, MWMS Client Communication Automation Framework, MWMS AIOS Lead Capture And Conversion Infrastructure Framework, MWMS High Ticket AIOS Client Acquisition And Trophy Client Framework, MWMS Client Intelligence Report Automation Framework, MWMS RAG Knowledge Base And Client Memory Infrastructure Framework, MWMS Data Extraction And Actor Infrastructure Framework, MWMS Meeting Intelligence And Action Extraction Framework, MWMS Source Visibility And Evidence Display Standard, MWMS Deep Search Quality And Observability Framework, MWMS AI Tool Permission And Access Framework, MWMS AI Automation Security And Risk Checklist, MWMS AI Output Validation Standard, MWMS Client Context Isolation And Privacy Boundary Standard
Purpose
The purpose of the MWMS AI Assisted Outreach And Sales Follow Up Automation Framework is to define how MWMS uses AI, structured prospect intelligence, approved context, automation, human review, and measurable sales workflows to support:
prospect discovery
lead enrichment
lead qualification
outreach research
personalised message drafting
inbound lead follow-up
cold outreach
warm outreach
proposal follow-up
voice-note style follow-up
appointment setting
reply handling
CRM progression
sales pipeline movement
reactivation
closeout
This framework exists because AI-assisted sales automation can create significant value when it helps MWMS or an AIBS client:
identify better prospects
understand genuine business problems
respond faster
send more relevant communication
maintain consistent follow-up
reduce forgotten leads
qualify opportunities
prepare proposals
book appointments
capture reply intelligence
improve sales process visibility
However, AI-assisted outreach can also create:
mass spam
fake personalisation
incorrect assumptions
creepy messaging
unsupported claims
privacy violations
deliverability damage
platform suspension
poor-fit appointments
incorrect CRM records
duplicate sends
ignored unsubscribe requests
reputational harm
This framework ensures that outreach and follow-up automation supports a governed sales process rather than becoming an uncontrolled message-generation system.
Core Doctrine
The MWMS doctrine is:
AI may research, enrich, score, prioritise, draft, classify, recommend, and coordinate.
AI must not automatically assume permission, buyer need, message accuracy, send authority, or sales readiness.
The complete governed sales communication lifecycle is:
Define Target Market
→ Define Buyer And Offer Fit
→ Discover Prospects
→ Verify Source And Permission Context
→ Enrich Prospect Record
→ Identify Evidence-Based Relevance
→ Score And Prioritise Prospect
→ Select Communication Channel
→ Determine Communication Authority
→ Draft Message
→ Validate Personalisation, Claims, Tone, And Compliance
→ Obtain Human Approval Where Required
→ Send Through Approved Channel
→ Confirm Delivery
→ Track Reply Or Non-Reply
→ Classify Response
→ Update CRM And Pipeline
→ Create Next Action
→ Stop, Continue, Escalate, Book, Park, Or Suppress
→ Measure Outcome
→ Improve Through Controlled Review
A sales automation is incomplete if it only generates a message.
It must also govern:
who should be contacted
why they are relevant
what evidence supports the personalisation
whether contact is appropriate
whether sending is authorised
what happens after the message
when follow-up stops
how replies are handled
how appointments are qualified
how CRM state is updated
what outcome is measured
Scope
This framework applies to:
AIBS prospecting
PPL prospecting
affiliate partnership outreach
sales inquiry follow-up
inbound lead follow-up
outbound lead generation
human-reviewed cold email
LinkedIn message drafting
WhatsApp sales follow-up where authorised
SMS follow-up where authorised
voice-note style follow-up
proposal follow-up
diagnostic report follow-up
appointment setting
lead reactivation
client acquisition
sales pipeline automation
CRM updates
reply classification
sales intelligence extraction
future client-facing outreach AIOS systems
This framework does not authorise mass automated outreach, unrestricted scraping, platform abuse, uncontrolled LinkedIn automation, automatic cold-email sending, or contact without compliance review.
Core Definition
AI-assisted outreach is the use of AI to help:
research prospects
identify relevance
draft communication
personalise messages
classify replies
recommend follow-up
coordinate sales actions
Sales follow-up automation is the governed use of automation to ensure qualified prospects receive appropriate next actions at the correct time.
MWMS Definition
The MWMS AI Assisted Outreach And Sales Follow Up Automation Framework is:
A governed Sales Brain system that converts authorised prospect and lead data into researched, prioritised, relevant, human-quality communication and controlled follow-up, while preserving compliance, permission, source evidence, human authority, CRM integrity, deliverability, suppression, and measurable sales outcomes.
Strategic Principle
The objective is not to send more messages.
The objective is to create more qualified sales conversations.
A smaller number of well-researched, relevant, properly timed messages may create more value than a large volume of generic AI outreach.
The MWMS standard is:
Quality Before Volume
Evidence Before Personalisation
Fit Before Contact
Draft Before Send
Permission Before Automation
Reply Quality Before Scale
Outcome Before Optimisation
The MWMS AI Assisted Outreach And Sales Follow Up Automation Model
Every outreach and follow-up system should operate across sixteen layers:
Market And Offer Layer
Lead Source And Authority Layer
Prospect Identity Layer
Prospect Intelligence Layer
Qualification And Prioritisation Layer
Permission, Suppression And Compliance Layer
Personalisation Evidence Layer
Message Architecture Layer
Channel And Delivery Layer
Human Review And Send Authority Layer
Follow Up Sequence Layer
Reply Classification Layer
Appointment And Sales Action Layer
CRM And Pipeline Layer
Measurement And Experimentation Layer
Learning And Governance Layer
- Market And Offer Layer
Outreach must begin with a defined market and offer.
The system should identify:
target market
buyer role
business type
problem addressed
desired outcome
offer
proof
delivery model
qualification criteria
disqualification criteria
approved claims
CTA
Rule
Do not begin prospecting until the target buyer and offer are clear.
Weak target definition produces generic outreach even when AI personalisation is used.
Market Questions
Ask:
Who has the problem?
How serious is the problem?
What evidence indicates the problem exists?
Can MWMS solve it?
Does the prospect have the ability to act?
Is the offer appropriate for the market?
What disqualifies the prospect?
- Lead Source And Authority Layer
Every prospect record must have a source.
Possible lead sources include:
inbound form
website inquiry
referral
event
newsletter response
CRM
approved business directory
public company website
authorised database
client-provided list
partner introduction
previous sales conversation
Research Brain output
Data Brain extraction workflow
Source fields should include:
Lead Source:
Source URL Or Record:
Capture Date:
Source Owner:
Permission Type:
Public Or Private:
Collection Method:
Permitted Use:
Compliance Status:
Rule
Access to contact information does not automatically create permission to send.
The source, intended use, jurisdiction, channel, and suppression status must be considered.
- Prospect Identity Layer
The system must identify the correct person, company, and relationship.
Identity fields may include:
Prospect ID:
Contact Name:
Company:
Role:
Email:
Phone:
Domain:
LinkedIn Profile:
Location:
Client Or Account:
Existing Relationship:
Previous Contact:
CRM Record:
Identity Confidence:
Possible identity statuses include:
Verified
Probable
Unverified
Duplicate
Conflicting
Rule
Do not create personalised outreach from an uncertain or conflicting identity without review.
- Prospect Intelligence Layer
Prospect intelligence should identify genuine context that may affect relevance.
Possible intelligence includes:
company service
company size
industry
location
recent business event
website condition
booking process
lead capture process
review profile
advertising activity
content activity
technology signals
customer complaints
job openings
growth signals
operational gaps
public offer
current CTA
known project
previous interaction
Intelligence must be classified as:
Verified Fact
Public Observation
Derived Signal
Estimate
AI Inference
Unverified Claim
Rule
A derived signal or AI inference must not be presented to the prospect as a confirmed fact.
Prospect Research Standard
Research should answer:
Why might this prospect be relevant?
What problem appears to exist?
What evidence supports that possibility?
What should not be assumed?
What could MWMS offer?
What is the safest low-friction next step?
- Qualification And Prioritisation Layer
Not every available contact should enter outreach.
Qualification may consider:
market fit
role fit
problem fit
offer fit
ability to pay
urgency
evidence quality
reachability
relationship strength
risk
permission
suppression status
timing
Prospect Priority Values
Priority One
Strong fit, credible evidence, appropriate timing, and clear next action.
Priority Two
Good fit but some evidence or timing remains uncertain.
Priority Three
Possible fit requiring more research.
Parked
Not ready, weak evidence, or timing unsuitable.
Rejected
Poor fit, prohibited, suppressed, or unacceptable risk.
Rule
Availability of contact data must not bypass prioritisation.
Prospect Qualification Record
Prospect:
Company:
Target Market Fit:
Role Fit:
Problem Evidence:
Offer Fit:
Ability To Act:
Timing:
Permission Context:
Risk:
Priority:
Qualification Status:
Recommended Next Action:
- Permission, Suppression And Compliance Layer
Before outreach, the system should check:
contact source
channel rules
jurisdiction
existing relationship
consent where required
legitimate business context where applicable
unsubscribe status
do-not-contact status
complaint history
client suppression list
internal suppression list
platform rules
frequency limits
personal-data use
Required suppression events include:
unsubscribe request
do-not-contact request
complaint
hard bounce
known invalid recipient
legal restriction
client exclusion
internal exclusion
Rule
Suppression must override automation.
No lead score, opportunity estimate, or campaign target may bypass a valid suppression state.
- Personalisation Evidence Layer
Personalisation must be based on real, relevant, appropriately obtained context.
Acceptable personalisation may include:
company service
public offer
recent company announcement
current website journey
publicly visible booking process
relevant review pattern
known inquiry
form answer
previous conversation
approved CRM record
verified business event
Personalisation should not include:
unrelated personal details
family information
private social information
sensitive personal information
unverified revenue estimates
invented compliments
fake familiarity
unsupported operational accusations
The Personalisation Evidence Record should include:
Prospect ID:
Personalisation Point:
Source:
Source Date:
Evidence Type:
Confidence:
Why Relevant:
Safe To Mention: Yes / No
Human Review Required: Yes / No
Rule
If MWMS cannot show the evidence behind a personalisation point, the point should not be used.
Personalisation Quality Standard
A personalisation point should be:
true
relevant
current
brief
respectful
non-invasive
connected to the offer
useful to the recipient
- Message Architecture Layer
An outreach message should have a clear structure.
Recommended First-Touch Structure
Relevant Context
Show why the communication is relevant.
Specific Observation
Mention one verified business observation or known need.
Problem Or Opportunity
Explain the potential issue or opportunity without overstating certainty.
Helpful Idea
Offer a useful thought, diagnostic, example, or question.
Low-Friction CTA
Suggest a simple next step.
A message should avoid:
generic introductions
long company biographies
fake praise
excessive personalisation
unverified claims
complex offers
multiple CTAs
false urgency
pressure
unsupported ROI
Message Types
First Touch
Introduces relevance and invites a low-friction response.
Value Add Follow-Up
Provides an additional useful observation.
Diagnostic Offer
Offers a short assessment, example, or resource.
Proposal Follow-Up
Clarifies value, objections, questions, or next steps after a proposal.
Appointment Follow-Up
Confirms or moves a qualified conversation toward booking.
Reactivation
Reopens a previous conversation using current and relevant context.
Closeout
Ends the sequence respectfully and gives the recipient control.
Rule
Each message must have one primary purpose.
- Channel And Delivery Layer
Possible channels include:
SMS
voice note
phone
website chat
CRM task
Channel selection should consider:
relationship
permission
buyer preference
message purpose
urgency
platform rules
deliverability
commercial appropriateness
Channel Rule
The existence of a channel does not mean it is appropriate.
The system should not automatically contact a prospect across several channels merely because contact details are available.
Cross-Channel Rule
Where several channels are used, the system must track:
which channel was used
when it was used
what was sent
whether the recipient replied
whether another channel is appropriate
suppression state
frequency
Rule
Cross-channel automation must not become cross-channel harassment.
- Human Review And Send Authority Layer
AI-generated outreach should begin in draft-only mode.
Possible authority levels include:
Research Only
AI may gather and organise prospect context.
Draft Only
AI may prepare a message but cannot send it.
Human Approved Send
AI or automation may send only after explicit human approval.
Approved Template Automation
An approved low-risk template may be sent under defined conditions.
Restricted
No drafting or sending is allowed.
Human review should check:
identity
fit
personalisation evidence
claim accuracy
tone
CTA
recipient
channel
permission
suppression
send timing
sequence state
Rule
Draft authority does not create send authority.
Send authority must be explicitly granted, scoped, logged, and reversible.
- Follow Up Sequence Layer
Follow-up must be controlled.
A follow-up sequence should define:
sequence purpose
eligible prospect status
channel
message count
timing
message purpose
stop conditions
reply handling
suppression handling
human review points
Sequence Example
Message One: Context And Helpful Idea
Introduce relevance, one observation, and one low-friction next step.
Message Two: Additional Value
Provide one useful observation, example, or clarification.
Message Three: Timing Or Objection Check
Ask whether timing, priority, or fit is the issue.
Message Four: Closeout
End respectfully and allow the prospect to re-engage later.
Rule
Follow-up should add value or clarify timing.
It should not repeat the same message with different wording.
Sequence Stop Conditions
Stop when:
recipient replies
recipient opts out
recipient complains
hard bounce occurs
identity becomes uncertain
prospect is disqualified
offer is no longer relevant
sales owner pauses the sequence
appointment is booked
human conversation begins
maximum follow-up count is reached
Rule
“Send until they reply” is prohibited.
- Reply Classification Layer
Replies must be classified before the next action.
Possible reply classes include:
Interested
Qualified interest or request for next step.
Question
The prospect requests information.
Objection
The prospect raises a concern.
Not Now
Timing is unsuitable but future interest may exist.
Not A Fit
The prospect or offer is unsuitable.
Referral
The recipient directs MWMS to another person.
Booked
An appointment or next step is confirmed.
Unsubscribe
The recipient requests no further communication.
Complaint
The recipient objects to the approach or data use.
Out Of Office
Temporary unavailability.
Bounce
Delivery failed.
Unknown
The reply cannot be classified confidently.
Rule
Reply classification must produce an action.
It must not merely add a label.
Reply Action Map
Interested
Assign sales owner and progress the opportunity.
Question
Prepare a grounded response or human handoff.
Objection
Route for objection-specific review.
Not Now
Park with an appropriate review date.
Not A Fit
Close or redirect.
Referral
Verify the referred contact before communication.
Booked
Confirm appointment and update CRM.
Unsubscribe
Suppress immediately.
Complaint
Stop outreach and escalate.
Out Of Office
Delay according to the stated return date where appropriate.
Bounce
Validate contact record and suppress invalid address.
Unknown
Route to human review.
- Appointment And Sales Action Layer
Outreach should create appropriate sales actions.
Possible actions include:
book diagnostic call
send approved resource
create proposal task
prepare audit
request additional information
route to sales owner
schedule follow-up
update qualification
park opportunity
close opportunity
suppress contact
Before booking an appointment, confirm:
prospect identity
interest
basic fit
purpose of meeting
timezone
availability
contact details
sales owner
calendar tool result
Rule
A booked appointment is not automatically a qualified opportunity.
Appointment quality should be measured.
- CRM And Pipeline Layer
Every active prospect should have a controlled CRM or pipeline state.
Possible states include:
Discovered
Researching
Qualified For Draft
Draft Ready
Waiting For Review
Approved For Send
Sent
Waiting For Reply
Replied
Qualification In Progress
Appointment Proposed
Booked
Proposal Sent
Follow-Up Due
Not Now
Parked
Suppressed
Rejected
Closed Won
Closed Lost
Rule
The CRM must reflect the actual current state.
The system must not mark a prospect as contacted, replied, booked, or qualified without supporting evidence.
Required CRM Fields
Prospect ID:
Company:
Contact:
Source:
Offer:
Priority:
Permission Context:
Personalisation Evidence:
Channel:
Last Communication:
Reply Class:
Owner:
Next Action:
Next Action Date:
Sequence Status:
Suppression Status:
Qualification Status:
Opportunity Value:
Outcome:
- Measurement And Experimentation Layer
Outreach should be measured across the full funnel.
Activity Metrics
prospects researched
prospects qualified
drafts created
messages approved
messages sent
follow-ups sent
Delivery Metrics
delivery rate
bounce rate
spam complaint rate
unsubscribe rate
Reply Metrics
reply rate
positive reply rate
question rate
objection rate
complaint rate
Qualification Metrics
qualified conversation rate
appointment proposal rate
booking rate
qualified booking rate
show rate
Commercial Metrics
proposal rate
close rate
revenue
sales-cycle length
cost per qualified conversation
cost per appointment
cost per client
Quality Metrics
personalisation accuracy
message relevance
human edit rate
reply-classification accuracy
CRM accuracy
follow-up timing accuracy
Pilot Before Scale Rule
Before scaling, MWMS should confirm:
messages are relevant
personalisation is accurate
deliverability is acceptable
complaints are low
reply quality is useful
bookings are qualified
CRM records are accurate
suppression works
human review remains manageable
Rule
Volume must not be increased merely because the workflow can send more.
- Learning And Governance Layer
The system should learn from:
human edits
reply patterns
objections
qualified conversations
disqualifications
complaints
bookings
no-shows
proposal outcomes
closed sales
failed sends
CRM corrections
Learning may improve:
target criteria
qualification
personalisation patterns
message structure
CTA
follow-up timing
objection handling
channel selection
sales handoff
Learning must not:
remove approval controls
weaken suppression
increase send volume automatically
change claims without approval
rewrite live sequences without review
Rule
Learning changes must be reviewed, versioned, tested, and approved.
Signal Triggered Outreach Standard
Outreach may be triggered by a meaningful business signal.
Possible signals include:
new inquiry
form submission
website visit where lawfully available
content download
proposal view
meeting attendance
new job posting
new product launch
new location
website change
review pattern change
funding announcement
lead-status change
previous follow-up date reached
Signal Trigger Requirements
Record:
Signal:
Source:
Signal Date:
Prospect:
Relevance:
Confidence:
Offer Relationship:
Permission Context:
Recommended Action:
Expiry:
Rule
A signal may improve timing.
It does not automatically prove buyer need or authorise contact.
Research To Message Standard
Research and message generation should remain separate stages.
Research Stage
Produce:
prospect facts
verified observations
possible needs
uncertainties
personalisation options
risks
qualification recommendation
Message Stage
Use only the approved and relevant research required for the communication.
Rule
The system should not expose the entire prospect research file in the message.
The message should use the smallest useful amount of context.
Prospect Intelligence Brief
Before important outreach, the system may create a Prospect Intelligence Brief.
Prospect Intelligence Brief Fields
Prospect:
Company:
Role:
Target Market Fit:
Known Relationship:
Verified Facts:
Public Observations:
Possible Problems:
Unverified Inferences:
Recent Signals:
Offer Fit:
Personalisation Options:
Risk:
Permission Context:
Recommended Channel:
Recommended Message Type:
Recommended CTA:
Human Review Notes:
Rule
The brief is an internal sales aid.
Inferences must not be presented externally as facts.
Inbound Lead Follow Up Standard
Inbound leads usually have stronger communication authority than cold prospects, but still require controlled handling.
Inbound workflow:
Lead Received
→ Confirm Source
→ Validate Contact
→ Capture Request
→ Classify Intent
→ Qualify
→ Determine Urgency
→ Draft Or Send Approved Response
→ Create CRM Record
→ Assign Owner
→ Schedule Follow-Up
→ Track Outcome
Inbound follow-up should consider:
form answers
requested service
source campaign
buyer urgency
budget indicators
location
fit
risk
Rule
An inbound form submission does not justify endless follow-up.
Frequency and suppression controls still apply.
Proposal Follow Up Standard
Proposal follow-up should use the actual proposal and sales context.
Inputs may include:
proposal date
proposal version
scope
price
problem
desired outcome
objections
decision process
stakeholders
next step
Proposal follow-up must not:
invent approval
change scope
change price
offer unauthorised discounts
create false urgency
misrepresent availability
Proposal Follow-Up States
Proposal Sent
Waiting For Confirmation
Questions Received
Revision Requested
Decision Pending
Follow-Up Due
Accepted
Rejected
Parked
Expired
Rule
Proposal follow-up should clarify and support the decision, not pressure the prospect.
Voice Note Style Follow Up Standard
Voice-note style communication may feel more personal.
It also carries authenticity and manipulation risk.
The system may:
draft a voice-note script
prepare a human recording brief
generate an audio draft where approved
It must not:
pretend an AI-generated voice message was personally recorded
clone a voice without authorisation
create false familiarity
use emotional pressure
hide automation where disclosure is required
Rule
Human-feeling communication must remain truthful.
Appointment Setting Assistant Standard
An appointment-setting assistant may:
confirm interest
ask approved qualification questions
offer available times
provide a booking link
book through an approved calendar tool
confirm the appointment
create a CRM event
It must not:
book poor-fit prospects merely to increase appointment volume
claim calendar success before confirmation
change commercial terms
answer unsupported questions
ignore timezone
continue after a human request
Rule
Appointment quality matters more than appointment count.
Sales Handoff Standard
When a prospect is ready for human sales contact, the handoff should include:
prospect
company
source
relationship
qualification
priority
verified context
personalisation used
messages sent
reply summary
questions
objections
offer
appointment details
risks
recommended next action
deadline
Rule
The sales owner should not need to reconstruct the entire history.
Outreach Sending Record
Every sent outreach communication should be traceable.
Record:
Communication ID:
Prospect ID:
Company:
Source:
Permission Type:
Offer:
Channel:
Message Type:
AI Draft:
Human Edited Version:
Personalisation Source:
Evidence Record:
CTA:
Compliance Flag:
Approval Status:
Sent Status:
Sent Date:
Provider Result:
Reply Status:
Outcome:
Next Action:
Sequence ID:
Suppression Status:
Last Updated:
Rule
AI-generated sales messages should be trackable.
Message Quality Standard
A message should pass this checklist before sending.
Relevant to recipient
Based on real context
Personalisation source recorded
Correct identity
Correct company
Correct offer
Appropriate channel
Short enough
Clear problem or opportunity
Safe claim
Human tone
Low pressure
Clear CTA
No fake urgency
No fake flattery
No unsupported ROI
No creepy personalisation
No excessive AI language
No private information
No prohibited commitment
Suppression checked
Rule
A polished message that fails identity, evidence, permission, or relevance checks is not ready.
Deliverability Standard
Outreach must protect sending reputation.
Monitor:
valid email rate
bounce rate
spam complaints
unsubscribe rate
domain health
send volume
message similarity
link use
attachment use
provider warnings
Deliverability controls may include:
email validation
volume caps
warming where appropriate
sending-domain separation
message variation based on genuine context
suppression
bounce handling
complaint escalation
Rule
Deliverability must not be protected through deceptive practices.
Reply And Follow Up State Machine
Each active sequence should have a defined state.
Possible states:
Not Started
Researching
Drafting
Waiting For Approval
Ready To Send
Sent
Waiting
Follow-Up Due
Reply Received
Human Review Required
Appointment Proposed
Booked
Parked
Closed
Suppressed
Failed
State transitions should be based on:
verified send result
elapsed time
reply
human decision
calendar result
suppression event
delivery failure
qualification outcome
Rule
The system must not send from an outdated or incorrect sequence state.
Idempotency And Duplicate Send Rule
Outbound workflows should prevent duplicate sends.
Controls may include:
communication ID
prospect ID
sequence ID
message-step ID
send-status check
provider message ID
time-window check
approved retry rule
Rule
A failed tool response must not trigger uncontrolled duplicate sending.
Human Review Approval Workflow
The human review workflow should show:
prospect
company
source
lead priority
permission context
personalisation evidence
message
claim risk
channel
sequence step
previous messages
reply state
recommended action
possible decisions:
Approve
Edit And Approve
Reject
Request More Research
Change Channel
Change CTA
Park
Suppress
Route To Sales Owner
Rule
Human approval should be informed, not a blind approve button.
Sales Automation Types
Type 1: Inbound Lead Follow Up System
Purpose:
respond quickly to inbound inquiries and create the correct next action.
Inputs:
form submission
source campaign
contact data
service interest
Outputs:
acknowledgement
qualification request
booking prompt
CRM update
follow-up task
Best Use:
AIBS leads
PPL leads
diagnostic funnels
Type 2: Lead Qualification System
Purpose:
score and classify leads before sales action.
Inputs:
form answers
CRM data
website behaviour
source campaign
company intelligence
Outputs:
lead score
qualification status
recommended next action
follow-up message
sales notes
Best Use:
AIBS
PPL
high-ticket offers
Type 3: Proposal Follow Up System
Purpose:
follow up after a proposal or diagnostic report.
Inputs:
proposal
proposal date
buyer problem
offer summary
objections
stakeholders
next step
Outputs:
follow-up email
reminder task
objection-specific reply
call-booking prompt
Best Use:
AIBS high-ticket sales
Type 4: Cold Outreach Drafting System
Purpose:
draft human-reviewed personalised outbound messages.
Inputs:
website URL
company data
public context
offer
target buyer
personalisation evidence
Outputs:
first email draft
LinkedIn message draft
follow-up sequence draft
prospect intelligence brief
Best Use:
carefully controlled AIBS prospecting
Type 5: Voice Note Style Follow Up System
Purpose:
create more human-feeling follow-up scripts or approved audio drafts.
Inputs:
lead context
form answers
sales stage
offer
previous interaction
Outputs:
voice-note script
approved audio draft
follow-up message
Best Use:
warm leads with human review
Type 6: Appointment Setting Assistant
Purpose:
move interested and qualified leads into booked appointments.
Inputs:
inbound message
lead details
qualification data
availability rules
timezone
Outputs:
booking link
appointment confirmation
CRM event
sales-owner notification
Best Use:
qualified warm leads
Type 7: Signal Triggered Outreach System
Purpose:
identify timely outreach opportunities based on meaningful business signals.
Inputs:
verified signal
source
prospect record
offer fit
permission context
Outputs:
research brief
priority update
draft message
review task
Best Use:
small-volume high-relevance prospecting
Type 8: Reply Intelligence And Routing System
Purpose:
classify replies and create the correct next action.
Inputs:
reply
thread context
prospect record
sequence state
Outputs:
reply class
risk
recommended response
CRM update
human handoff
suppression event where required
Best Use:
inbound and outbound sales workflows
Build Path
Stage 1: Define Offer And Target
Document:
offer
buyer
problem
qualification
disqualification
proof
approved claims
CTA
Stage 2: Define Lead Source
Document:
source
authority
permission
capture method
required fields
suppression rules
Stage 3: Create Prospect Intelligence Process
Define:
research questions
source requirements
evidence types
inference labels
personalisation evidence
Stage 4: Define Qualification
Create:
fit rules
priority
risk
permission
timing
routing
Stage 5: Define Message Architecture
Create:
message types
tone
CTA
claim limits
channel rules
Stage 6: Set Human Review
Define:
draft authority
send authority
review fields
approval status
exceptions
Stage 7: Build Follow Up States
Define:
sequence
timing
stop conditions
reply classes
next actions
Stage 8: Connect CRM
Define:
stages
owners
dates
communication records
suppression
outcomes
Stage 9: Test Failure Paths
Test:
wrong identity
wrong company
duplicate send
bad source
invalid address
unsubscribe
complaint
negative reply
unknown reply
tool failure
calendar failure
wrong timezone
wrong sequence state
Stage 10: Pilot Before Scale
Use a small, controlled prospect group.
Review:
message quality
reply quality
deliverability
complaints
qualification
booking quality
human workload
Stage 11: Improve Through Evidence
Improve only from:
verified outcomes
human edits
reply evidence
sales feedback
CRM accuracy
deliverability data
Launch Readiness Checklist
Before launching outreach or follow-up automation, confirm:
offer is defined
target buyer is defined
qualification is defined
disqualification is defined
lead source is recorded
source authority is reviewed
identity checks exist
prospect research method exists
personalisation evidence is stored
inference is labelled
permission is reviewed
suppression is working
channel is appropriate
message structure is approved
claims are controlled
CTA is clear
human review exists
send authority is explicit
sequence steps are defined
stop conditions exist
reply classes exist
reply actions exist
CRM stages exist
owners exist
next-action dates exist
duplicate-send controls exist
deliverability monitoring exists
complaint handling exists
unsubscribe handling exists
calendar confirmation exists where used
measurement exists
pilot size is controlled
scaling conditions are defined
Failure Modes
Failure Mode 1: Mass AI Spam
Large numbers of AI-generated messages are sent without relevance.
Correction
Reduce volume, improve fit, and require evidence-based personalisation.
Failure Mode 2: Fake Personalisation
The message contains generic observations presented as personalised research.
Correction
Store and review the personalisation evidence.
Failure Mode 3: Creepy Personalisation
The message uses unnecessary personal details.
Correction
Use business-relevant context only.
Failure Mode 4: Unverified Prospect Identity
The message is addressed to the wrong person or company.
Correction
Verify identity and domain before drafting or sending.
Failure Mode 5: Enrichment Treated As Fact
AI-generated enrichment is presented as confirmed truth.
Correction
Label enrichment as fact, estimate, inference, or unverified.
Failure Mode 6: Every Record Is Contacted
All records from a database or scraper enter a sequence.
Correction
Apply qualification and prioritisation first.
Failure Mode 7: Draft Permission Becomes Send Permission
A drafting system begins sending automatically.
Correction
Separate tool authority and require explicit send approval.
Failure Mode 8: No Suppression
Unsubscribe or complaint records remain active.
Correction
Create immediate suppression and block future sends.
Failure Mode 9: Follow-Up Never Stops
The sequence continues until a reply.
Correction
Define maximum steps and stop conditions.
Failure Mode 10: Same Message Repeated
Follow-ups add no new value.
Correction
Give each sequence step a separate purpose.
Failure Mode 11: Unsupported ROI Claim
The message promises savings or revenue without evidence.
Correction
Use approved and qualified claims only.
Failure Mode 12: Wrong Channel
The system uses a channel that is inappropriate or unauthorised.
Correction
Apply channel selection and permission rules.
Failure Mode 13: Cross-Channel Harassment
The prospect receives repeated messages across several platforms.
Correction
Track all communication in one sequence state.
Failure Mode 14: Duplicate Send
A retry or workflow error sends the same communication twice.
Correction
Use communication, sequence, step, and provider IDs.
Failure Mode 15: Reply Is Ignored
The sequence continues after a response.
Correction
Stop the sequence immediately when a reply is detected.
Failure Mode 16: Reply Is Misclassified
A complaint or unsubscribe is treated as interest.
Correction
Use conservative classification and human review for uncertainty.
Failure Mode 17: Bad Appointments
The system books unqualified prospects.
Correction
Add minimum qualification before booking.
Failure Mode 18: CRM State Is False
The CRM reports activity or progress that did not occur.
Correction
Require supporting send, reply, booking, or decision evidence.
Failure Mode 19: Proposal Follow-Up Changes Terms
The AI offers a discount or scope change without authority.
Correction
Detect commercial commitments and require human approval.
Failure Mode 20: Voice Note Pretends To Be Personal
AI-generated audio is presented as a human-recorded personal message.
Correction
Require truthful use and disclosure where appropriate.
Failure Mode 21: Signal Treated As Need
A business event is treated as proof the prospect wants the offer.
Correction
Use the signal only to improve timing and relevance.
Failure Mode 22: Research Is Exposed In Message
The outreach message includes excessive or invasive research detail.
Correction
Use the smallest safe personalisation point.
Failure Mode 23: Scaling Before Reply Quality
Volume increases because delivery works, even though replies are poor.
Correction
Scale only after positive reply and qualification evidence.
Failure Mode 24: Human Review Becomes Mechanical
Reviewers approve without examining identity, evidence, or risk.
Correction
Display review evidence and require meaningful decisions.
Governance Responsibilities
Sales Brain
Owns:
sales process
qualification
message purpose
follow-up strategy
reply handling
pipeline stage
sales handoff
commercial outcome
AIBS Brain
Owns:
AIBS prospect strategy
client acquisition offers
AIOS sales use cases
client-facing sales automation packaging
Research Brain
Supports:
prospect research
company intelligence
signal verification
source quality
market context
Data Brain
Owns:
prospect records
identity
deduplication
CRM fields
sequence states
communication records
measurement integrity
Automation Brain
Owns:
workflow triggers
schedules
delivery actions
retry controls
tool confirmation
duplicate-send protection
CRM orchestration
Compliance And Risk Brains
Own:
channel rules
jurisdiction review
consent
suppression
personal-data use
platform risk
claim risk
complaint handling
Content Brain
Supports:
human-quality message language
tone
voice
clarity
message templates
Customer Brain
Supports:
relationship continuity
buyer experience
preference handling
respectful communication
Experimentation Brain
Owns:
controlled message tests
CTA tests
sequence tests
channel tests
pilot design
statistical interpretation
HeadOffice
Owns:
cross-Brain conflicts
high-risk exceptions
authority changes
strategic review
scale approval where material
Future AI Employee Ideas
These AI Employee ideas are parked candidates only.
Outreach Personalisation Analyst
Primary Brain: Sales Brain / Research Brain
Purpose:
Reviews public and provided lead context to create accurate, relevant, non-creepy personalisation points.
Prospect Intelligence Analyst
Primary Brain: Research Brain / Sales Brain
Purpose:
Creates source-backed prospect intelligence briefs and separates facts, observations, estimates, and inference.
Sales Follow Up Coordinator
Primary Brain: Sales Brain / Automation Brain
Purpose:
Tracks follow-up timing, pipeline stage, next action, and sequence status.
Lead Qualification Analyst
Primary Brain: Sales Brain / Data Brain
Purpose:
Scores leads based on problem fit, offer fit, permission, urgency, budget indicators, and risk.
AI Proposal Drafting Assistant
Primary Brain: Sales Brain / AIBS Brain
Purpose:
Creates proposal drafts from diagnostic findings, scope, deliverables, risks, and value logic for human approval.
Outreach Compliance Reviewer
Primary Brain: Compliance Brain / Risk Brain
Purpose:
Reviews email, SMS, WhatsApp, LinkedIn, affiliate, and PPL outreach for consent, disclosure, unsubscribe, platform, and claim risk.
Deliverability Monitor
Primary Brain: Data Brain / Risk Brain
Purpose:
Monitors bounce rate, spam complaints, domain health, send volume, suppression lists, and campaign risk signals.
Reply Classification Analyst
Primary Brain: Sales Brain / Data Brain
Purpose:
Classifies replies and recommends the governed next action.
Appointment Setting Assistant
Primary Brain: Sales Brain / Automation Brain
Purpose:
Helps qualified leads move into booked calls while avoiding poor-fit calendar clutter.
Sales Sequence Optimiser
Primary Brain: Sales Brain / Experimentation Brain
Purpose:
Reviews reply rates, booking rates, message quality, CTA performance, and sequence outcomes.
Voice Follow Up Governance Reviewer
Primary Brain: Compliance Brain / Risk Brain
Purpose:
Reviews AI-generated audio or voice-note workflows for consent, authenticity, disclosure, and manipulation risk.
Sales CRM Integrity Reviewer
Primary Brain: Data Brain / Sales Brain
Purpose:
Checks whether pipeline stages, communication states, appointments, next actions, and outcomes match supporting evidence.
Drift Protection
This framework protects MWMS from:
mass AI spam
fake personalisation
unreviewed cold outreach
weak lead qualification
over-automated sales
unsupported sales claims
poor deliverability
ignored unsubscribe rules
platform suspension
creepy personalisation
voice-note manipulation
LinkedIn automation abuse
proposal overpromising
CRM chaos
sales messages hidden in chat history
follow-up without strategy
scaling before reply quality is known
contacting every record returned by a database
skipping prioritisation because data is available
treating enrichment as proof of buyer need
draft permission becoming send permission
signals being treated as proof of buyer intent
duplicate sends
cross-channel harassment
unqualified appointment volume
reply classifications without next actions
AI inference being presented as prospect fact
Drift Signals
Watch for:
“Let’s send this to everyone.”
“AI can personalise it enough.”
“No need to review these.”
“The list is public, so it is fine.”
“Let’s automate all LinkedIn messages.”
“We do not need unsubscribe wording.”
“The voice note will feel personal.”
“We can promise ROI.”
“Send until they reply.”
“The email looks good enough.”
“The CRM can be updated later.”
“Volume will fix it.”
“Let’s scale before testing replies.”
“The database found them, so they are qualified.”
“We can send because the draft is ready.”
“More enriched records means a better campaign.”
“The signal means they need us.”
“Use everything we found in the message.”
“Bookings are all that matter.”
“The workflow said it sent, so it must have sent.”
Rule
When these drift signals appear, return to fit, source evidence, qualification, permission, human review, send confirmation, suppression, reply quality, and outcome measurement.
Implementation Boundary
This framework is an architecture and operating standard.
It does not authorise immediate development of:
outreach agents
mass-email systems
LinkedIn automation
scrapers
CRM automations
voice-note generation
calendar agents
lead databases
client outreach products
Before implementation, HeadOffice or the authorised Brain must create a scoped build brief defining:
use case
target market
offer
lead source
permission context
data fields
qualification
personalisation evidence
channel
message type
human review point
approval status
sending method
suppression rules
CRM stage
follow-up schedule
reply classes
measurement metrics
stop criteria
Rule
No outreach automation should be built without a defined sales process and compliance boundary.
Strategic Summary
AI-assisted outreach and sales follow-up are valuable when they create:
better research
better prioritisation
more relevant communication
faster response
consistent follow-up
clearer sales records
better appointment quality
stronger sales intelligence
The value does not come from message volume.
The value comes from connecting:
market fit
prospect evidence
qualification
permission
personalisation
human-quality communication
controlled sending
reply handling
CRM state
measurable outcomes
The strongest starting systems are:
Inbound Lead Follow Up
Human-Reviewed Cold Outreach Drafting
Proposal Follow Up
Reply Intelligence And Routing
Qualified Appointment Setting
Signal Triggered Prospect Research
These systems should begin with narrow scope, draft-first operation, human review, controlled pilots, and visible evidence.
Final Standard
The MWMS final standard is:
AI-assisted outreach must begin with a defined target market and offer, use authorised and source-visible prospect data, qualify before contact, personalise from real evidence, separate drafting from sending, respect suppression, manage replies, maintain accurate CRM state, and measure sales outcomes before scale.
A valid outreach and follow-up system must define:
target market
buyer
offer
lead source
source authority
identity
prospect intelligence
qualification
priority
permission
suppression
personalisation evidence
message architecture
channel
human review
send authority
delivery confirmation
follow-up sequence
stop conditions
reply classes
reply actions
appointment rules
CRM stage
owner
next action
measurement
pilot criteria
scale criteria
That is the MWMS AI Assisted Outreach And Sales Follow Up Automation Standard.
MWMS System Change Log
Version: v1.2
Date: 2026-06-21
Author: HeadOffice
Change
Updated the MWMS AI Assisted Outreach And Sales Follow Up Automation Framework from v1.1 to v1.2 using the later AI Automations by Jack material covering:
AI client intelligence
automated company research
RAG-assisted communication
prospect intelligence
personalised outreach
sales follow-up
reply handling
appointment setting
CRM automation
proposal follow-up
voice-note style communication
signal-triggered workflows
Expanded the existing outreach model into sixteen governed layers:
• Market And Offer Layer
• Lead Source And Authority Layer
• Prospect Identity Layer
• Prospect Intelligence Layer
• Qualification And Prioritisation Layer
• Permission, Suppression And Compliance Layer
• Personalisation Evidence Layer
• Message Architecture Layer
• Channel And Delivery Layer
• Human Review And Send Authority Layer
• Follow Up Sequence Layer
• Reply Classification Layer
• Appointment And Sales Action Layer
• CRM And Pipeline Layer
• Measurement And Experimentation Layer
• Learning And Governance Layer
Added new standards covering:
• prospect identity
• prospect intelligence classification
• source-backed personalisation evidence
• signal-triggered outreach
• research-to-message separation
• Prospect Intelligence Briefs
• cross-channel communication control
• controlled reply classification
• reply-to-action mapping
• appointment quality
• sales handoff context
• outreach communication records
• follow-up state machines
• idempotency and duplicate-send prevention
• proposal follow-up state
• CRM evidence integrity
Added explicit doctrine that:
• contact-data availability does not create permission
• prospect research must distinguish verified facts from inference
• signals improve timing but do not prove buyer need
• drafting and sending are separate permissions
• cross-channel access must not become cross-channel harassment
• reply classification must produce a next action
• appointment quality matters more than appointment count
• CRM state must match supporting evidence
Added new Sales Automation Types:
• Signal Triggered Outreach System
• Reply Intelligence And Routing System
Added AI Employee candidates:
• Prospect Intelligence Analyst
• Sales CRM Integrity Reviewer
Expanded failure handling to cover:
• incorrect prospect identity
• enrichment treated as fact
• signals treated as buyer intent
• duplicate sends
• cross-channel harassment
• reply misclassification
• poor-quality appointments
• false CRM states
• exposed research detail
• mechanical human approval
Change Impact Declaration
This update strengthens the existing Sales Brain framework without changing its ownership, parent page, or authority.
Sales Brain remains the primary owner.
AIBS Brain, PPL Brain, and Affiliate Brain may use the governed outreach architecture within their own commercial workflows.
Research Brain supports prospect intelligence.
Data Brain supports identity, state, deduplication, and measurement.
Automation Brain may execute approved workflow steps but does not determine permission, qualification, or send authority.
The update does not authorise:
• unrestricted prospect scraping
• mass automated sending
• uncontrolled LinkedIn automation
• automatic cold outreach by default
• voice cloning without authority
• use of sensitive personal information
• suppression bypass
• unreviewed commercial commitments
Pages Created
• None
Pages Updated
• MWMS AI Assisted Outreach And Sales Follow Up Automation Framework updated from v1.1 to v1.2
Pages Deprecated
• None
Standalone Pages Not Created
The following standalone pages were not created because their durable intelligence is governed within this updated framework:
• MWMS Signal Triggered Outreach Framework
• MWMS Prospect Intelligence Brief Framework
• MWMS Research To Outreach Framework
• MWMS Reply Intelligence And Routing Framework
• MWMS Cross Channel Sales Follow Up Framework
• MWMS Sales CRM State Machine Framework
• MWMS AI Appointment Setting Framework
• MWMS Proposal Follow Up Automation Framework
• MWMS Outreach Duplicate Send Prevention Standard
• MWMS Personalisation Evidence Standard
Registries Requiring Update
• MCR Page Registry
• Sales Brain Page Registry
• MCR Copy Map where the framework version is recorded
• MWMS Course Absorption Decision Registry
Canon Version Update Required
No immediate Sales Brain Canon version change is required unless the Canon directly records framework versions or contains outreach rules that conflict with v1.2.
The new prospect-intelligence, evidence-based personalisation, reply-state, duplicate-send, appointment-quality, and CRM-integrity controls should be included during the next scheduled Sales Brain Canon alignment review.
Change Log Entry Required
Yes.
The v1.2 update must be recorded in:
• MWMS System Change Log
• MCR Page Registry change history where applicable
• Sales Brain Page Registry change history where applicable
• MWMS Course Absorption Decision Registry
Strategic Absorption Result
The later AI Automations by Jack material concerning company research, client intelligence, RAG-assisted communication, personalised outreach, response handling, appointment setting, and sales automation has been absorbed into the existing MWMS AI Assisted Outreach And Sales Follow Up Automation Framework.
The absorption preserves the durable sales-automation intelligence while rejecting:
• mass-message volume as strategy
• contact-data access as permission
• AI inference as prospect truth
• fake personalisation
• signals being treated as proof of buyer intent
• automatic send authority
• endless follow-up
• cross-channel harassment
• appointment volume without qualification
• CRM progression without evidence
• proposal changes without authority
• uncontrolled voice-note impersonation
The resulting v1.2 framework establishes that MWMS outreach and follow-up must be:
• market-led
• offer-aligned
• source-visible
• identity-aware
• evidence-personalised
• qualification-controlled
• permission-safe
• suppression-enforced
• human-reviewed
• send-authorised
• reply-responsive
• CRM-traceable
• appointment-quality focused
• outcome-measured
END OF FULL FILE OUTPUT