Client
PropEdge Realty
Industry
Real Estate
Tech Stack
PythonHubSpot APIOpenAITwilioReactPostgreSQL
The Challenge
What We Were Up Against
PropEdge's 120 agents were spending 60% of their day on cold outreach that converted at under 2%. They had 8,000+ leads in a CRM but no way to prioritise — everyone worked the list from the top.
Our Solution
How We Solved It
We built an AI lead scoring engine trained on 3 years of closed deals, integrated it with their CRM, and automated all nurture touchpoints for low-score leads — freeing agents to focus exclusively on the 91%-accurate high-intent prospects.
01
CRM Data Archaeology
Cleaned and enriched 8,000 leads with property search behaviour, engagement history, and demographic overlays.
02
Lead Scoring Model
Trained a gradient boosting model on 3 years of closed deals to predict purchase intent — 91% accuracy on holdout data.
03
Automation Workflows
Built 14 automated nurture sequences in HubSpot: SMS, email, and WhatsApp at key intent signals.
04
Agent Prioritisation UI
Custom CRM dashboard surfacing each agent's top 10 leads daily, with conversation starters and property matches.
05
Feedback Loop
Agent call outcomes feed back into the model weekly — accuracy improved from 87% to 91% in 3 months.
The Impact
Measurable Results
Deal close rate increased 2.3× within 6 months
Agents reclaimed 60% of cold-lead time for high-value activity
Lead scoring accuracy reached 91% on new enquiries
Revenue grew 38% without adding a single agent
Average days-to-close reduced from 48 to 29
"Our top agents used to work every lead the same way. Now the AI tells them who to call first — and they're closing 2× as many deals in the same hours. It's transformational."
Brian Holloway
CEO, PropEdge Realty
Key Outcomes
Deals closed2.3×
Time on cold leads−60%
Lead scoring accuracy91%
Revenue increase38%
