Tripled qualified property leads with an AI sales agent

Dhaka Realty Group, one of the larger real estate developers in Bangladesh, was generating a steady volume of inbound enquiries from Facebook ads, website forms, and property portals like bproperty.com — but their sales team of 12 consultants could not keep up. Most enquiries arrived outside business hours, and leads were being responded to 6–18 hours later, by which point many had already engaged competitors. Their legacy Zoho CRM had no automation, no lead scoring, and no WhatsApp integration, so consultants were manually copying lead data between Facebook Lead Ads spreadsheets, the website form submissions, and WhatsApp chat windows. Duplicate leads, missed follow-ups, and zero visibility on conversion rates made it impossible to forecast revenue. Worse, their combined Google and Meta ad spend had grown to ৳4.5 lakh per month with no clear attribution — they were paying for clicks that turned into conversations that never closed. Management knew they had a leaky funnel but could not pinpoint where, and the consultants were burning out trying to chase cold leads alongside genuinely interested buyers. The marketing team was also retargeting the same audiences with diminishing returns, unable to identify which segments actually converted, so an estimated 60% of ad budget was being wasted on leads who would never buy. The Managing Director had been pitched by three other agencies promising "AI solutions" but was skeptical after a previous vendor delivered a chatbot that broke every time a visitor asked an off-script question.

We deployed a bilingual (Bangla/English) AI chat agent on their website that engages every visitor in under 3 seconds, qualifies them against a 9-question property-buyer profile (budget, location preference, ready flat vs. under-construction, handover timeline, financing method), and routes hot leads directly to the right property consultant. The same AI agent was extended to the WhatsApp Business API so Facebook Lead Ads enquiries automatically receive a personalized WhatsApp message within 60 seconds. We then integrated everything into a fresh GoHighLevel CRM with lead scoring, automated follow-up sequences (5 touchpoints over 14 days via WhatsApp, email, and SMS), and a real-time dashboard for the sales manager. A site-visit booking flow lets the AI propose available time slots and sync them to the assigned consultant's Google Calendar. Every conversation is logged, transcribed, and scored, so management can identify high-performing consultants and coach underperformers. A custom UTM-tracking layer finally closed the attribution gap between ad spend and closed deals. We also implemented a sentiment and intent classifier that scores every conversation in real-time, surfacing only sales-ready leads to consultants via WhatsApp push notifications — meaning consultants now spend 80% of their day on closing conversations instead of prospecting. A weekly automated performance report is generated for the Managing Director showing lead-source ROI, consultant leaderboard, and pipeline forecast. The entire stack was deployed in 3 weeks with zero changes to the existing website — the AI agent was embedded as a floating widget plus a dedicated WhatsApp Business number.