The Problem
Real estate agencies generating leads from Facebook, TikTok, and WhatsApp had no single system to manage them. Follow-ups were manual and inconsistent, agents missed messages, response times stretched into hours, and there was no shared view of where a lead stood. WhatsApp — the client's primary channel for buyer and seller communication — ran through personal numbers with zero reporting or automation, so there was no visibility into delivery, read, or reply rates, and no way to run structured follow-up campaigns.
The Solution
I designed and built PropFlow CRM end-to-end: an ASP.NET Web API backend integrated directly with the official WhatsApp Business API via Meta (no third-party BSP layer), paired with Google Firestore for real-time chat sync and a broadcast/automation engine for template-based messaging. Every inquiry — regardless of source — lands in one dashboard, gets auto-assigned to an agent, and is tracked from first message to close.
Key Features
WhatsApp Business API integration — sends template, text, and media messages through Meta's official API, with webhook-based real-time delivery and read/reply tracking, so nothing depends on unofficial WhatsApp automation tools.
Broadcast automation & analytics — bulk campaign sends with per-broadcast metrics (sent, delivered, read, replied, failed) and custom date-range filtering, turning WhatsApp into a measurable channel instead of a black box.
Lead & inquiry management — a centralized dashboard with automatic agent assignment, lead status tracking, and structured follow-up workflows.
Real-time chat — live WhatsApp conversations synced instantly through Google Firestore, with a full one-to-one conversation history per client.
Template library & personalization — reusable, dynamic-variable message templates for fast, on-brand outreach at scale.
Technical Highlights
Built on ASP.NET (MVC + Web API) with 100+ optimized REST endpoints and an async repository pattern for better concurrency under load. SQL Server handles the system of record, tuned with indexing and stored-procedure optimization to keep reporting and broadcast queries fast as contact volume grows, while Firestore handles the low-latency real-time layer — a split that kept the chat experience instant without giving up relational integrity for leads, agents, and campaigns.
Results
Automating WhatsApp follow-ups replaced ad hoc manual replies with consistent, on-time responses and freed agents to focus on closing rather than chasing inboxes. Broadcast analytics gave the business its first real visibility into WhatsApp performance, and centralized lead assignment meant inquiries stopped falling through the cracks.
FAQ
Why integrate directly with Meta's WhatsApp Business API instead of a third-party automation tool?
Third-party WhatsApp automation tools run outside Meta's official terms and can get numbers banned with no warning. A direct Meta Business API integration means message delivery, templates, and webhooks are all officially supported, so the client's WhatsApp number stays safe as usage scales up.
How does the system track whether a WhatsApp message was actually delivered or read?
Meta's API sends delivery and read-receipt webhooks in real time, which the system captures per message and rolls up into per-broadcast analytics — sent, delivered, read, replied, and failed counts — instead of agents guessing whether a message actually reached someone.
Why use both SQL Server and Firestore instead of just one database?
SQL Server holds the relational system of record — leads, agents, campaigns, assignments — where data integrity and reporting queries matter most. Firestore handles the real-time chat layer, which needs to feel instant. Splitting the two meant the live chat experience didn't have to compromise on the reliability of the CRM's core data.
How does lead assignment work when inquiries come from three different channels?
Every inbound inquiry — whether it started on Facebook, TikTok, or WhatsApp — lands in one centralized dashboard and is automatically assigned to an agent, so leads get the same follow-up treatment and visibility regardless of which platform they originated from.
What kind of scale does the backend handle?
The API layer has 100+ optimized REST endpoints built with an async repository pattern for concurrency under load, with SQL Server tuned through indexing and stored-procedure optimization specifically so broadcast and reporting queries stay fast as contact volume grows.
