Three portals, one conversation graph

Definition

Duplicate lead merge is the job of detecting the same mobile number (or email) across portal dumps, web, Meta, and CP intake, collapsing to one Buyers & Tenants record and one Booking, and continuing the omnichannel thread without a second RM calling.

Speed-to-lead dies when two closers race

Portals sell the same listing. Two teams call. The buyer opts out. CoPulse: search Customers by phone before creating. Workflows cannot invent a marketplace-CRM suppression list (that gap exists between marketplace allocations and CRM). For own-database rows, humans merge. Do not promise adaptive matching.

Omnichannel sequence

Any capture hop can create a twin. Merge, then resume WhatsApp qualify → voice visit.

How it runs in CoPulse

Mostly human plus inbox search. Workflows can notify on same-phone create if you built that check. Conversation agent continues after merge. Autonomous must not double-send both rows. Do not use autonomous tasks as the inbound qualifier. Conversation agents, Growth drips, and autonomous list sends are three systems with no shared stop-when-booked contract. WhatsApp live catalog is Gupshup, not an invented Meta Cloud hop. Phone is Plivo; email is SES. Wallet-only pricing: preview the Organization wallet before a blast. Organization and Project — never a shared inbox that vanishes when a broker leaves.

Operator setup

Train ops to search. Optional properties for all sources. Wallet wasted on double templates is the tell.

99acres Monday, Housing Tuesday, website Wednesday

Three Bookings, one number. Ops merges Tuesday night. Qualifier already asked budget on WhatsApp; Housing template is cancelled. Visit Scheduler keeps Saturday 11. Click-path: Customers duplicate view, inbox, Growth stop on the extra recipient if a campaign was queued.

Marketplace vs CRM suppression is not shipped — do not mix unrevealed marketplace allocations into this story. Illustrative funnel only after one graph remains. Pricing: wasted duplicate sends already hit the wallet; merging stops the bleeding.

Operators assemble this job on the Project. Click-path remains Customers search → merge → Agents → Router single owner → Growth lists cleanup → wallet preview so you notice double sends. Keep the quotable definition in the brief for the team: Duplicate lead merge is the job of detecting the same mobile number (or email) across portal dumps, web, Meta, and CP intake, collapsing to one Buyers & Tenants record and one Booking, and continuing the omnichannel thread without a second RM calling. Do not use autonomous tasks as the inbound qualifier. Conversation agents, Growth drips, and autonomous list sends are three systems with no shared stop-when-booked contract. While you implement, keep the parent playbooks directory, New enquiry, WhatsApp overlay, Brokers, Agent Studio, Growth campaigns, and wallet-only pricing in the same runbook. Do not invent Meta Cloud WhatsApp if the catalog is Gupshup. Do not promise adaptive nurture. Human handoff stays in the omnichannel inbox. Appointments and calendars are the visit truth when this job books time. Library templates are required outside the 24-hour WhatsApp window. Workflows fire on pipeline stage change only when you attach them under Agents → Workflows. The agent router keeps the wrong bot off this thread. Knowledge bases hold PDFs you uploaded. Preview the Organization wallet before a blast. Property managers should not paste sales token stages onto leasing jobs. The Duplicate leads playbook never forks into a WhatsApp-only or voice-only child URL. Funnel percentages on some leaves are library template estimates from Site-visit booking blitz, New project launch teaser, or Cold-buyer win-back — not an SLA and not Insights. If this job mentions money, the Organization wallet is the meter. If it mentions a Saturday slot, Visit Caller and T-24h templates are in-product; cab dispatch is not. That is the honest operating manual for this URL.

Channel sequence (one job, not extra URLs)

Dedupe is a control-plane job with channel side-effects. There is no nested playbook-by-channel path.

Illustrative funnel (library template estimates)

Double-counting reached is how fake funnels appear — merge first. These percentages come from CoPulse library playbook templates (Site-visit booking blitz, New project launch teaser, or Cold-buyer win-back). They are not your project's Insights chart.