540
Paid
6.5% of total base ↑
782
Lapsed
Was premium → now free ↑
7,040
Free
Trial + Free combined ↑
8,362
Total Agents
Since Aug 2023 →
User Tier Breakdown
Paid
540
Lapsed
782
Free
7,040
Hot ≤7d
1,108
Warm 8-30d
685
At Risk 31-90d
295
Dormant >90d
6,274
Action Summary
459
Hot Upgrade Leads
Score ≥9 — pitch now
→
1,031
Warm Leads
Score 5-8 — nurture
→
782
Lapsed Winback
Were premium — re-engage
→
9
Critical Churn Risk
Expiring ≤7 days
→
0
Behavioral Risk
Disengaged paid users
→
255
Testimonial Candidates
Paid + active + sold
→
388
Power Users
Listed AND co-marketed
→
Monthly Signups — Click a Bar to Open Signup Analytics
ⓘ
V2 weighted score 0–20 across 4 dimensions: listing volume, co-marketing depth, recency, and activity signals. Hot ≥9 → call now. Warm 5–8 → nurture. Cold → activate first.
459
Hot Score ≥6
Click to see all →
1,031
Warm Score 3–5
Click to see all →
5,550
Cold Score 0–2
Activation needed first
Score Signal Reference
| Signal | Points | Rationale |
|---|---|---|
| listing_total > 0 | +2 | Has content on platform |
| listing_total > 5 | +2 | Active lister — more to gain from premium |
| comarketing_request_dt not null | +2 | Tried to collaborate |
| listing_publish_dt not null | +1 | Published live — serious intent |
| listing_chg_price_dt not null | +1 | Active seller |
| profile_comarketing_search_dt not null | +1 | Seeking partners |
| comarketing_download_dt not null | +1 | Downloaded data — power intent |
| days_since_active ≤ 7 | +2 | Active now — peak conversion window |
⚡
782 agents already paid once. Re-engagement cost is lower than new acquisition.
782
Total Lapsed
Click to see all →
~180
High Priority (score ≥5)
Winback immediately
1.45×
Lapsed vs Active Paid
More lapsed than paying
Lapsed Score Signals
| Signal | Points | Meaning |
|---|---|---|
| listing_total > 10 | +3 | Heavy lister — platform was core to their work |
| listing_sold_dt not null | +2 | Closed deals — proven results |
| days_since_active ≤ 30 | +2 | Still coming back |
| comarketing_request_dt not null | +1 | Used collaboration |
| share_listing_to_client_dt not null | +1 | Used platform to engage clients |
9
Critical
Call NOW →
32
High
This week →
21
Medium
Re-engage →
478
Healthy
>60 days remaining
Action Queue
🔴
Critical (9): Personal call today. Renewal incentive. Do not let lapse.
🟡
High (32): Renewal reminder this week with listing stats.
🔵
Medium (21): Re-engagement email + feature highlight.
Paid Recency
⚠
V2 Behavioral Scoring: These paid agents score Critical (≥7/10) based on inactivity + low engagement — regardless of expiry date. They are at true churn risk even if their subscription has months remaining. Proactive outreach now costs far less than a renewal campaign later.
0
Behavioral Critical
Score ≥7 →
0
Inactive >90 days
Highest inactivity signal
0
Zero Platform Usage
No listings AND no co-mkt
Why Behavioral Risk Matters
🔴
Expiry-date alone misses this: An agent with 300 days left who has been inactive for 90 days is a higher true churn risk than one expiring in 20 days who published a listing yesterday.
🟡
What to do: Outreach focused on value realisation — help them get their first listing up, or show them the co-marketing feature. These agents paid but never extracted value.
🔵
Scoring dimensions: Inactivity 0–4pts · Low engagement 0–3pts · Expiry urgency 0–3pts. Score ≥7 = Critical, ≥5 = High, ≥3 = Medium.
Churn Score Breakdown (Paid)
All Audiences
Loading…
ⓘ
Criteria: Paid + listing_total >5 + active + (sold OR co-marketed)
255
Total Candidates
Review candidates →
47%
of All Paid Users
Have provable results
3
Outreach Tiers
Video / Feature / Volume
Outreach Strategy
1
Tier 1 — Sold + Active ≤7 days: Best for video testimonial. Highest credibility.
2
Tier 2 — Co-marketed + Sold: Best for co-marketing feature case study.
3
Tier 3 — High Volume + Active: listing_total >20 this month. Productivity angle.
8,362
Registered
100% — Day 0
2,088
Ever Active
25% used platform
1,326
Activated
Listed or co-marketed →
388
Power Users
Both actions →
Breakdown
388
Power Users
Listed AND co-marketed
→
338
Listing Only
No co-marketing yet
600
Co-marketing Only
Uses others listings
7,036
Not Activated
Never listed or co-marketed
Gaps
🔴
84.1% never activated. Onboarding is the #1 product problem.
⚡
600 co-marketed without listing. Upsell opportunity.
ⓘ
Target: 40%+ of signups publish first listing within 7 days.
Active Features — Paid vs Free
| Feature | Users | All % | Paid % | Free % | Gap | Signal |
|---|
Awaiting Data
Kartel
0 of 16 events
Shop
0 of 11 events
PMCoin
0 of 10 events
Affiliate
0 of 4 events
Tracking Gaps
🔴
comarketing_accept_dt — ALL NULL. 988 requests, 0 accepts tracked.
🟡
prospect_search_dt — NULL.
🟡
notification_read_dt — NULL.
🔵
listing_new_form_dt — NULL.
| Agency | Agents ▾ | Paid | Paid% | Lapsed | Hot Leads | Listings | Co-Mkt | Action |
|---|
i
ADVISORY ONLY Registration-based branch/group intelligence for internal validation. Not used for campaigns, WhatsApp, recommendations, or automation.
Agent Tiers
Paid Paid Agent
Agent currently on an active PREMIUM subscription. The highest-value segment — protect these at all costs.
subscription == 'PREMIUM'
Lapsed Lapsed Agent
Was previously on a PREMIUM plan but has since downgraded or expired back to FREE. They understood the value enough to pay once — highest-intent winback targets.
subscription_type == 'PREMIUM' AND subscription == 'FREE'
Free Free Agent
Never paid, or signed up on a trial that didn't convert. Includes both FREE and TRIAL subscription_type. The largest segment and the top of the upgrade funnel.
subscription == 'FREE' AND subscription_type IN ('FREE', 'TRIAL')
Recency Bands
Hot Hot (≤7 days)
Agent was meaningfully active on the platform within the last 7 days. Best window to reach out — highest response rate.
days_since_active <= 7
Warm Warm (8–30 days)
Active within the last month. Still engaged — good for nurture sequences and follow-ups.
days_since_active between 8 and 30
At Risk At Risk (31–90 days)
Haven't been active for over a month. May be slipping away — a targeted re-engagement prompt can recover these.
days_since_active between 31 and 90
Dormant Dormant (>90 days)
No meaningful activity in over 3 months. Bulk of the free base. Requires activation intervention before upsell is realistic.
days_since_active > 90, or active_dt is null
Activity Signals & Data Fields
active_dt
Date of the agent's last meaningful platform action — e.g. publishing a listing, requesting co-marketing, using search. This is the primary recency signal.
⚠ Not all logins update active_dt. Use this over login_dt.
login_dt
Date of last login. Unreliable as a recency measure because persistent browser cookies keep sessions alive without the agent actually returning. Used only as a fallback when active_dt is null.
⚠ Do not use for recency scoring — inflated by cookie sessions.
Ever Active
Agents who have triggered at least one meaningful action (active_dt is not null). Only 25% of registered agents ever reach this state — a key onboarding health metric.
active_dt IS NOT NULL → 2,088 agents (25%)
listing_total
Cumulative all-time count of listings associated with this agent. Includes active, sold, and expired listings. Used for activation and scoring — more reliable than listing date fields.
comarketing_total
Cumulative all-time count of co-marketing activities. Agents who co-market tend to be platform-committed — strong upgrade signal.
days_since_active
Number of days between the snapshot date and active_dt. Drives all recency band classifications. Null if agent has never been active.
snapshot_date − active_dt (fallback: − login_dt)
days_to_renewal
Days remaining until subscription_expiry_dt. Used exclusively for Paid agents to determine churn risk urgency. Negative means already expired.
subscription_expiry_dt − snapshot_date
REN Level
Real Estate Negotiator license tier — a proxy for agent experience and seniority. REN is the standard registration category under BOVAEA in Malaysia.
Activation Funnel
Activated
Agent has completed at least one core value action: listed a property OR used co-marketing. This is the first meaningful product milestone after signup.
listing_total > 0 OR comarketing_total > 0
Power User
Agent has done both: listed properties AND used co-marketing. These agents are deeply embedded in the platform workflow — highest retention and LTV.
listing_total > 0 AND comarketing_total > 0
Listing Only
Has published listings but never used co-marketing. One action away from becoming a Power User — prime upsell target for co-marketing features.
listing_total > 0 AND comarketing_total == 0
Co-marketing Only
Has used co-marketing (leveraging other agents' listings) but has no own listings. May be early-stage agents or those building a book before listing independently.
comarketing_total > 0 AND listing_total == 0
Not Activated
Never listed and never co-marketed. These agents signed up but never discovered the core product loop. Onboarding is the primary lever here, not sales.
listing_total == 0 AND comarketing_total == 0
Scoring Models
Upgrade Score (Free agents, max 20 — V2)
V2 weighted behavioral score predicting likelihood to convert to Paid. 4 dimensions: listing volume, co-marketing depth, recency, and activity signals. Higher = call first.
Listing volume (0–7): 0→0, 1–2→1, 3–10→3, 11–30→5, 31+→7
Co-mkt volume (0–5): 0→0, 1–5→2, 6–20→4, 21+→5
Recency (0–4): ≤7d→4, ≤14d→3, ≤30d→2, ≤90d→1
Activity signals (0–4, 1pt each):
listing_publish_dt · listing_chg_price_dt
share_listing_to_client_dt · comarketing search/download
Co-mkt volume (0–5): 0→0, 1–5→2, 6–20→4, 21+→5
Recency (0–4): ≤7d→4, ≤14d→3, ≤30d→2, ≤90d→1
Activity signals (0–4, 1pt each):
listing_publish_dt · listing_chg_price_dt
share_listing_to_client_dt · comarketing search/download
Hot Lead ≥9 · Warm Lead 5–8 · Cold <5
Lapsed Score (Lapsed agents, max 15 — V2)
V2 weighted winback priority score. Higher = stronger re-subscription candidate. Considers content investment, proven ROI, and continued platform activity.
Listing depth (0–5): 0–2→0, 3–10→2, 11–30→3, 31+→5
Had a sale (0–3): listing_sold_dt set→+3
Recency (0–4): ≤7d→4, ≤30d→3, ≤60d→2, ≤90d→1
Co-mkt activity (0–2): request+1, download+1
Sharing (0–1): share_listing_to_client_dt set→+1
Had a sale (0–3): listing_sold_dt set→+3
Recency (0–4): ≤7d→4, ≤30d→3, ≤60d→2, ≤90d→1
Co-mkt activity (0–2): request+1, download+1
Sharing (0–1): share_listing_to_client_dt set→+1
Churn Risk (Paid agents, V2 behavioral)
V2 behavioral churn score (0–10). Goes beyond expiry date — a paid user inactive for 90 days is a higher true churn risk than one expiring soon but actively publishing listings.
Inactivity signal (0–4): >90d→4, >60d→3, >30d→2, >14d→1
Low engagement (0–3): no listings+co-mkt→3, no listings→1
Expiry urgency (0–3): ≤7d→3, ≤30d→2, ≤60d→1
Critical ≥7 · High ≥5 · Medium ≥3 · Healthy <3
Low engagement (0–3): no listings+co-mkt→3, no listings→1
Expiry urgency (0–3): ≤7d→3, ≤30d→2, ≤60d→1
Critical ≥7 · High ≥5 · Medium ≥3 · Healthy <3
Call Critical today. High this week. Medium in nurture.
Testimonial Candidate
Paid agents who are good candidates to feature in marketing materials — verified active, high listing volume, and have demonstrated platform value through a sale or co-marketing activity.
tier == 'Paid'
AND listing_total > 5
AND active_dt IS NOT NULL
AND (listing_sold_dt IS NOT NULL OR comarketing_request_dt IS NOT NULL)
AND listing_total > 5
AND active_dt IS NOT NULL
AND (listing_sold_dt IS NOT NULL OR comarketing_request_dt IS NOT NULL)
Dashboard Concepts
Snapshot Date
The date the raw agent data was extracted from the database. All recency calculations (days_since_active, days_to_renewal) are relative to this date — not today.
Hot Lead
A Free agent with Upgrade Score ≥9 (max 20, V2 behavioral scoring). Scored across 4 weighted dimensions: listing volume, co-marketing depth, recency, and activity signals. Highest-priority sales calls each week.
Warm Lead
A Free agent with Upgrade Score 5–8. Showing strong intent but not yet hot — typically 1–2 actions away. Suitable for automated nurture sequences (email, WhatsApp).
Lapsed Winback
Former paid agents now on Free. Sorted by Lapsed Score. These have demonstrated willingness to pay — re-engagement cost is lower than acquiring a new paid user.
Feature Adoption
Breakdown of which platform features (Feeds, Listings, Co-marketing, Profile, Share, Prospect, Notification) have been used by agents, and what % of paid vs free users use each. Low adoption = onboarding opportunity.
Dashboard Overview
🔒 Internal use only
This dashboard is for internal intelligence review by PropMall operators, sales, media, and management teams. It is not a customer-facing tool and should not be shared outside the team.
🎯 What it helps you identify
The dashboard surfaces five types of opportunity: Upgrade leads (free agents ready to convert), Lapsed winback (former paid agents to re-engage), Paid retention (churn risk), Testimonial candidates (agents with strong story angles), and Agency/branch intelligence (structural and behavioral signals by office).
⚠️ Labels are decision support
Tiers, scores, and badges are decision aids — not final instructions. They surface patterns from snapshot data. Operators should verify context before acting. No outreach is triggered automatically from this dashboard.
🕑 Snapshot-based data
All data reflects a specific snapshot export date shown in the sidebar. Recency figures (days active, days to renewal) are relative to that snapshot date, not today. Rebuild the dashboard after each new data export to get current figures.
Testimonial Candidates Guide
Tier 1 Recently Successful Active Agents
Paid agents who have a sold listing result AND were active on PropMall within the last 7 days. Best for fresh success-story interviews — the outcome is recent and easy to recall. Interview angle: what helped the deal move, what they did differently.
Tier 2 Co-marketing Success Candidates
Paid agents with a sold result AND measurable co-marketing usage. Good for collaboration and network-effect testimonials — they can explain PropMall as a shared inventory tool. Interview angle: who they collaborated with, what outcome followed.
Tier 3 High-volume Productive Agents
Paid agents with more than 20 listings and recent activity (within 30 days). Good for workflow and productivity testimonials — they can describe repeatable PropMall usage habits. Interview angle: listing routine, how they stay organized.
📋 Candidate Detail panel
Click any row in the candidate table to open the detail drawer. It shows the agent's behavioral signals, the system-suggested story angle, and the recommended interview direction. Use this to prepare before a call — it is read-only and does not record anything.
📱 Missing mobile
Some agent records do not have a mobile number in the system. These are flagged in red. Before preparing outreach for these agents, verify their contact details through another internal source (CRM, registration records, or direct account lookup).
✅ How to use the list
Use the tier filters, listing range, and co-marketing range to narrow the candidate pool. Select candidates using checkboxes to build a shortlist. The selection count is shown in the summary bar. Export and outreach must be prepared manually outside this tool.
Agency V2 Explorer Guide
📑 Registration Group
A cluster of agency registration numbers that share the same canonical agency name. For example, all E(1)1766/x registrations belong to the Maxxan group. The group row shows aggregated signals across all branches.
🏠 HQ Node
The primary registration number for an agency (no slash suffix, e.g. E(1)1766). Represents the head office. Agents registered under the HQ number are counted at this node.
🏢 Branch Node
A numbered sub-registration (e.g. E(1)1766/3). Represents a physical branch office. When a known branch label has been curated, it shows the branch name (e.g. "Ipoh (Greentown)"). When not yet curated, it shows "Unknown Branch".
❓ Orphan Node
An agent registration number that exists in the agent data but has no matching record in the agency registration map. These agents are real but their branch affiliation is not yet confirmed. Shown with a warning mapping chip.
🔹 Unknown Branch
A branch node whose label has not yet been curated in the registration map. The registration number is valid and the agents are real — only the human-readable branch name is missing. This is a data completeness gap, not an error.
🔴 Advisory Health
A signal-based health label (Healthy, Watch, Concern, Critical) for each branch node. Derived from agent tier mix, activation rate, listing density, and churn risk across agents in that node. It is advisory — not a definitive diagnosis.
🔍 Identity Confidence
Indicates how reliably the system has matched agents to this registration node. High confidence means most agents carry matching registration data. Low confidence means the node is inferred from partial signals and may contain misassigned agents.
Branch / Agent Drilldown Guide
🔍 Opening the drilldown
In the Agency V2 Explorer, expand any agency group by clicking "Inspect nodes". Each branch node row has a 👤 Agents button. Click it to open the agent panel scoped to that branch. Close with the ✕ button or press Escape.
🎗 Filters and sorting
The panel has six filter dimensions: free-text search, tier (Paid / Free / Lapsed), activation status, priority / tags, listing count, and co-marketing count. All filters apply instantly. The table shows only agents assigned to the selected branch registration number.
📋 Filtered snapshot
The panel header shows a live count of agents matching the current filters. Use this to quickly assess branch composition — for example, filter to "Paid" to see how many paid agents are at a specific office.
✅ Selection is for review only
Checkboxes in the agent table allow you to mark agents for review. The selection count is shown in the summary bar. No outreach, export, or record change is triggered from this panel. Selection state is lost when the panel is closed.
🚫 No outreach from this view
The drilldown panel is an operational preview only. It has no WhatsApp, email, CRM write, or export capability. Any follow-up actions based on what you see here must be prepared and executed outside the dashboard using your standard outreach tools.
Note: If the panel shows 0 agents for a branch, it means no agent records in the current data export carry that branch registration number. This may indicate the branch is new, the agents registered under a different number, or the registration data is incomplete for that office.
Data Confidence & Limitations
🏢 Branch labels may be incomplete
Not all branch registration numbers have a curated human-readable label yet. These appear as "Unknown Branch" in the explorer. The registration data is valid — only the display name is missing. Curation is ongoing.
🕑 Snapshot freshness matters
Metrics like "Days Active" and "Days to Renewal" are calculated from the snapshot date shown in the sidebar — not from today. An agent who was "active 3 days ago" at snapshot time may be 2 weeks old by the time you act. Rebuild the dashboard after each new export.
📈 Advisory labels are not final truth
Scores, health labels, persona tags, and tier placements are computed from available signals. They guide attention — they do not replace judgment. An agent classified as "churn risk" may already have renewed; a "hot lead" may have cancelled. Verify before committing outreach time.
📱 Mobile numbers are not validated
Mobile numbers shown in the dashboard come directly from the agent registration record. They are not validated for accuracy or reachability. Always confirm contact details before an outreach campaign, especially for high-priority agents.
🔗 Identity confidence is signal-based
Agent-to-branch assignment is based on registration number matching. Where registration numbers are missing or inconsistent in agent records, the system makes best-effort assignments. Low-confidence nodes may contain agents who belong to a different branch.
⚖️ Operators should verify before action
Before any outreach or operational decision based on dashboard signals, verify through at least one additional source — CRM, account records, or direct contact. Dashboard intelligence is a starting point, not a final authority.
Operator SOP
1
Review the cohort. Navigate to the relevant dashboard section (Upgrade Leads, Lapsed Winback, Testimonials, Agency V2 Explorer, etc.). Use the filters to scope by tier, recency, agency, or score range to identify the priority pool for this session.
2
Inspect candidate or agent detail. Click into individual rows to open the detail drawer or candidate panel. Review behavioral signals, suggested story angle or action type, and any risk flags. Use this to decide whether the agent warrants action this cycle.
3
Shortlist. Use checkboxes to mark agents you intend to follow up on. The selection summary shows count, agency distribution, and tier breakdown. You can filter further before selecting to ensure you are shortlisting the right agents.
4
Verify contact information. For shortlisted agents, confirm mobile numbers and other contact details using your CRM or registration records before preparing outreach. Agents with "Missing mobile" in the dashboard must be resolved before outreach can proceed.
5
Prepare outreach manually. Draft messages, calls, or email sequences outside the dashboard using your standard tools (WhatsApp, email platform, CRM). The dashboard does not send any communication — it only surfaces who to contact and why.
6
Record feedback externally. Log call outcomes, interview notes, or conversion results in your CRM or shared team tracker. The dashboard does not record feedback at this stage. This ensures the team's institutional knowledge is preserved outside the intelligence tool.
Reminder: This dashboard is a read-only intelligence surface. No actions, no exports, no outreach, and no record changes are made from within it. All follow-up is external and manual.
Audience Lists
—
Saved audience lists
Import CSV or analytics export →
Active Campaigns
—
Campaigns in progress
View campaigns →
Open Conversations
—
Awaiting operator reply
Open console →
Templates Ready
—
Meta-approved templates
View library →
AI Drafts Pending
—
Suggested replies for review
Human approval required →
Follow-ups Outstanding
—
No reply sent yet
Conversations needing action
Outreach Workflow
01
Analytics Export
Behavioral signals identify agents worth contacting
→
02
Audience List
Operator selects segment, imports as named audience
→
03
Campaign
Assign Meta-approved template + audience
→
04
Template Send
First message via WhatsApp Business API
→
05
Recipient Reply
Inbound reply opens 24-hour window
→
06
Human / AI
Operator replies; AI suggests, never auto-sends
→
07
Outcome
Conversation resolved, outcome logged
Quick Access
Audience Lists
Import and manage contact lists from analytics or CSV
Campaigns
Create and track WhatsApp outreach campaigns
Campaign Operations
Lifecycle segment readiness and governed sequence planning
Conversation Console
3-panel chat interface with agent intelligence context
Template Library
Meta-approved message templates for campaign sends
AI Assistance
How AI suggests replies — human approval always required
Compliance
Meta policy, opt-out rules, and PDPA requirements
Recent Campaigns
All →
Loading…
Open Conversations
Console →
Ahmad Razif
IQI Realty · Interested, boleh cerita lagi?
09:32
Siti Hajar 2
REN Properties · Ok boleh, saya nak tanya...
Yesterday
Faridah Idris
IQI Realty · Serius ke? Boleh explain lagi
Sunday
Phase 3 (data model) → Phase 5 (live replies). Placeholder data shown.
Compliance Snapshot
Full rules →
⚠
Meta template approval required before first outreach. Only pre-approved templates can initiate a conversation.
⚠
Opt-out is permanent. Any STOP reply or button opt-out must immediately prevent all further messages to that contact.
ⓘ
PDPA-aware handling. Agent contact data is personal data under Malaysian law. Do not share outside the operator team.
All Audience Lists
Loading…
Total
—
All campaigns
Active
—
In progress
Draft
—
Not launched
Review
—
Awaiting operator approval
All Campaigns
Loading…
Needs Reply
0
0 overdue
Follow Up Due
0
Due now or overdue
Stale
0
No recent activity
Converted
0
Recently converted
Unassigned
0
No operator assigned
Action Center
Deterministic focus map — no AI ranking
Loading action center...
Total Conversations
0
0 active
Updated Today
0
Manual activity
Avg Response
—
Inbound to outbound
Queue Risk
0
Overdue + stale
Campaign Attribution
| Campaign | Targets | Convs | Replies | Interested | Converted | Risk |
|---|---|---|---|---|---|---|
| Loading… | ||||||
Agency Responsiveness
| Agency | Convs | Reply Rate | Converted | Due |
|---|---|---|---|---|
| Loading… | ||||
Loading…
Select a conversation
Outreach Console — Phase 3
Select a conversation from the left panel.
Local conversation record only — WhatsApp API not connected yet.
Operator Copilot
Select a conversation to load the copilot workspace.
PropMall Ops
—
propmall_ops templates
Project Sales
—
project_sales templates
Approved
—
Meta approved
Draft / Pending
—
Awaiting Meta approval
Loading templates…
AI Role in WhatsApp Outreach
🤖
AI suggests replies only. Claude generates a suggested response based on conversation history and agent intelligence context. The suggestion appears as a draft — it is never sent automatically.
📈
AI summarises the conversation. For longer threads, AI can produce a brief summary of what was discussed and the contact’s apparent intent — displayed in the intelligence panel for operator context.
🎯
AI recommends next action. Based on reply intent (interested / not interested / question / complaint), AI can suggest: call booking, send pricing info, escalate, or resolve. Operator confirms or overrides.
👤
Human approval is always required. The operator must review, edit if needed, and explicitly click Send. There is no auto-send path at any phase of this system.
🌐
Language follows the recipient. If the agent replies in Bahasa Malaysia, AI drafts in Bahasa Malaysia. Language is not changed without an explicit operator instruction.
🚫
AI cannot make pricing or promotion commitments. The AI prompt explicitly prohibits specific pricing claims, limited-time offers, or commitments not pre-approved by management. Operators are responsible for all commitments made in the conversation.
Rules & Requirements
⚠
Meta template approval required. Only approved templates can initiate a WhatsApp conversation. Unapproved messages violate WhatsApp Business Policy and risk number suspension. Submit via Meta Business Manager before any campaign launch.
⚠
Opt-out must be respected immediately. Any “STOP”, “Unsubscribe”, or button opt-out must prevent all further messages to that contact. The system enforces this automatically — do not manually re-add opted-out contacts to new campaigns.
ⓘ
Human review required before every send. No message — template or free-form — is sent without a deliberate operator action. This applies to AI-suggested messages as well. Review content and context before sending.
ⓘ
No bulk spam blasting. This tool is designed for targeted, intelligence-driven outreach. Audience lists must reflect genuine behavioral signals, not entire agent databases. High send volumes with low engagement degrade the phone number’s quality rating.
ⓘ
PDPA-aware contact handling. Agent phone numbers and personal data are personal data under Malaysia’s Personal Data Protection Act. Contacts are stored for legitimate business purposes only. Opt-out triggers data minimization. Do not share contact lists outside the operator team.