Skip to content
Ricardo Brossard
Projects

CRM Lexinton

A real estate agency's inquiries, in one system with no manual entry.

Real estate agency in Buenos AiresIn production, 10 users

The system's WhatsApp inbox, with the team's conversations split into valuations and sales
inquiries processed with no manual entry
3,500+
of last month's portal inquiries, linked to their property automatically
98%
audited status changes
6,805
users on the team
10
A client's inquiry arrives on its own from ZonaProp; her record shows the linked property and the assigned agent; a viewing is booked and lands in the agent's Google Calendar; the client replies on WhatsApp and the scheduled automated messages are canceled; and the dashboard shows how long the team takes to reply. With sample data.A client's inquiry arrives on its own from ZonaProp; her record shows the linked property and the assigned agent; a viewing is booked and lands in the agent's Google Calendar; the client replies on WhatsApp and the scheduled automated messages are canceled; and the dashboard shows how long the team takes to reply. With sample data.A client's inquiry arrives on its own from ZonaProp; her record shows the linked property and the assigned agent; a viewing is booked and lands in the agent's Google Calendar; the client replies on WhatsApp and the scheduled automated messages are canceled; and the dashboard shows how long the team takes to reply. With sample data.A client's inquiry arrives on its own from ZonaProp; her record shows the linked property and the assigned agent; a viewing is booked and lands in the agent's Google Calendar; the client replies on WhatsApp and the scheduled automated messages are canceled; and the dashboard shows how long the team takes to reply. With sample data.
From the listing portal to a booked viewing with no manual entry, and the automated messages go quiet when the client replies.

The problem

Inquiries from buyers and tenants arrived every day through several channels (the listing portals, the Facebook campaigns, the website) and ended up in a shared email inbox. Each one depended on someone copying it by hand into a spreadsheet, and the owner had no way to know how many came in, which ones were answered or who was handling each client.

What I built

A system that gathers all those inquiries in one place without anyone entering anything: each one arrives on its own, already tied to the property it asks about and assigned to the agent in charge, and the team follows up from there, including the agency’s WhatsApp.

To make that work, the system connects to Tokko Broker, the software where the agency manages its properties and its agents, and mirrors its inventory with prices, developments and removed listings. Every 15 minutes it reads the inbox, recognizes which portal each inquiry comes from, detects whether that person had asked before (also when the same inquiry arrives directly from the portal and forwarded by Tokko) and links it to its property by the listing code or, if there is none, by the address.

WhatsApp is integrated through Meta’s official API: every conversation sticks to the client’s record, and the automated follow-up messages stop on their own when the client replies, so a person takes over. Viewings are booked in each agent’s Google Calendar, and for the first time the owner sees the business in numbers: inquiries by channel, the status of each one and who is handling them.

How it works

  1. 016 channelsportals, Meta, web
  2. 02Email readerevery 15 minutes
  3. 03Detects duplicateson four levels
  4. 04Links the propertycode or address
  5. 04Tokko Brokerinventory, read-only
  6. 05Assigns the agentper property
  7. 06Management systemstatuses, dashboard
  8. 07WhatsAppinbox and 7 flows
  9. 07Google Calendarviewings per agent

Results

  • 207 valuations and 186 viewings booked

    Valuation requests come in on their own from the forms, and viewings go to each agent's Google Calendar.

  • 7 automated flows

    WhatsApp follow-up messages that are set up without touching code and stop when the client replies.

  • 100 unit tests and 33 end-to-end tests

    Every new module is validated in a private beta, with real data, before it goes to production.

Stack

Application
Next.js 14ReactTypeScriptTailwind CSS
Data and backend
SupabasePostgreSQLRow Level SecurityRealtimeEdge Functions on Deno
Integrations
WhatsApp Cloud APIGmail APIGoogle Calendar APITokko Broker API
Quality and deployment
VitestPlaywrightVercel
Dashboard with today's inquiries, the last 7 days and the all-time total, and the weekly trend