Our Work
Real Projects, Honest Results
Every case study documents what was actually built and what was actually delivered — no fabricated metrics, no stock-photo mockups, no vanity numbers.
Mustaraka Properties
Full-stack AI-ready real estate platform — chatbot trained on live listings, automated lead pipeline, and self-hosted admin CRM.
The Challenge
Mustaraka Properties, a premium real estate agency in Dhaka, was launching its digital presence from scratch. They needed more than a website — they needed an integrated system that could capture leads from Facebook Ads and the website simultaneously, notify agents instantly, handle initial enquiries via AI at any hour, and give the team complete control over listings, leads, and content without depending on any third-party SaaS tool that would become a recurring cost.
What We Built
We built a complete Next.js platform (App Router, TypeScript, Tailwind CSS) with a property listings database and a built-in AI chatbot powered by Google Gemini, trained on live listing data so it can answer real property enquiries. Every lead from the website contact form or a Meta lead ad triggers an n8n workflow that fires instant WhatsApp and Telegram alerts to agents — pre-formatted with the lead's name, phone, enquiry type, and property interest. A full admin CRM gives the team property CRUD with image upload, blog management, and lead tracking with status updates, all behind an authenticated admin panel. The entire stack is self-hosted on a VPS behind Traefik with automated SSL — no Vercel, no Zapier, no external CRM seats.
Technology Stack
Results & Outcomes
Platform Status
Live from launch day
AI Chatbot
Handling property enquiries 24/7
Lead Pipeline
Every submission → instant WhatsApp + Telegram alert
SaaS Dependency
Zero — Zapier, HubSpot, and CRM seats replaced
Lead Conversion Rate
[ To be updated after 60 days live ]
Nova Digital Store
69-node WhatsApp receipt-verification workflow — manual approval replaced with automated validation and duplicate detection.
The Challenge
A high-volume digital products store was receiving hundreds of payment receipt screenshots via WhatsApp every day. Customers would send images of bank transfers or mobile payment confirmations to complete their purchases. Staff were manually inspecting each image, matching reference numbers against transaction records, and approving or rejecting orders — a process that was slow, error-prone under load, and completely unscalable during promotions and peak demand periods.
What We Built
We designed and built a 69-node n8n workflow on the client's VPS. When a customer sends a payment receipt to the WhatsApp Business number (via Evolution API), the workflow automatically extracts the transaction amount and reference number using OCR, cross-references the record against a Google Sheets order ledger, runs duplicate-detection logic to catch repeat submissions from the same transaction, and sends the customer a confirmation or rejection message — all without staff involvement. A separate alert notifies staff only for edge cases that fall outside the automated rules and require manual review. Full workflow documentation and handover was delivered in Spanish to match the client's working language.
Technology Stack
Results & Outcomes
Automated Processing
Standard transactions handled without staff intervention
Duplicate Detection
Double-submission attempts caught and flagged automatically
Edge-Case Routing
Ambiguous receipts routed to staff with full context — no case dropped
Documentation
Full workflow docs and handover delivered in Spanish
Case Study · Coming Soon
Your Business
Your project could be the next case study here.
Start Your ProjectYour Project
Ready to Build Something Real?
Tell us what you want to automate. We'll scope it clearly, build it properly, and document everything — so your business owns it forever.
Book a Free Consultation