The complete picture — including the complications
Benchmarks are useful. Case studies are more useful. This article documents the operational transformation of a mid-size Gurgaon residential brokerage — 380–450 leads per month, selling inventory across three premium projects on the Dwarka Expressway corridor and one on Sohna Road — from a human-only calling operation to AI-augmented calling. The numbers are from an actual 6-month operational period. The brokerage's identity is not disclosed at their request, but every metric cited is verifiable from their CRM reporting.
The story includes the month-2 qualification script problem, the initial closer team resistance, and the after-hours lead recovery that turned out to be the single largest revenue driver. The complete picture, including the complications, is more useful than a curated success narrative.
The Brokerage: Baseline Profile
A channel partner brokerage in Gurgaon. 400–450 leads/month from MagicBricks, 99acres, Housing.com, and developer direct referrals. Three projects on Dwarka Expressway (Sectors 102, 106, 113), one on Sohna Road. Calling team: 4 BDRs + 1 team lead, 9 AM–7 PM Monday to Saturday on Sell.Do CRM. Average commission per booking: ₹3,20,000. Monthly marketing spend: ₹1,85,000.
| Metric | Baseline Value |
|---|---|
| Contact rate | 44% |
| Qualified leads per month | 48 |
| Site visits per month | 17 |
| Bookings per month | 3.1 |
| Monthly revenue | ₹9,92,000 |
| Calling team cost (fully loaded) | ₹2,85,000 |
| Total monthly cost (calling + marketing) | ₹4,70,000 |
| CAC | ₹1,51,613 |
| Net monthly contribution | ₹5,22,000 |
The brokerage was profitable but not scaling. 17 site visits per month from 420 leads represented a 4% lead-to-visit conversion rate. The principal's assessment: "We know we're losing leads. I can see it in the CRM — leads marked 'not connected' that never got followed up because the team was overwhelmed with the ones who did pick up."
Month 1: Deployment and Baseline Reset
Deployment timeline: 4.5 weeks from contract signing to live calls.
Weeks 1–2 — CRM integration. Sell.Do API configuration, lead intake webhook setup, field mapping for 14 qualification data points. One technical issue: the Sohna Road project had a different lead source format from the Dwarka Expressway projects — required additional routing logic. Resolved in 3 days.
Week 3 — Script development. Four qualification scripts — one per project — covering: budget range, BHK preference, possession timeline sensitivity, financing status, location priority within the corridor (Sectors 102–106 vs. 108–113 for Dwarka Expressway), and decision-maker availability. The Sohna Road script included an additional question on SPR vs. Sohna connectivity preference, which proved to be a high-signal qualifier for that project's buyer profile.
Week 4 — Testing and QA. 50 test calls with live leads. Manual review of CRM records against transcripts. Finding: the budget field was populating inconsistently when buyers gave ranges ("1.5 to 1.8 crore" was extracting as ₹1.5 crore, losing the upper range). Fixed with updated NLU configuration.
| Metric | Baseline | Month 1 | Change |
|---|---|---|---|
| Contact rate | 44% | 79% | +35 pp |
| Qualified leads | 48 | 89 | +85% |
| Site visits | 17 | 29 | +71% |
| Bookings | 3.1 | 5.2 | +68% |
| Revenue | ₹9,92,000 | ₹16,64,000 | +68% |
Month 1 showed a significant jump — but not the full potential. Contact rate at 79% was below the 84–92% target because the calling window was still restricted to 9 AM–8 PM, and the follow-up sequencing for non-connected leads had not been fully activated.
Month 2: The Qualification Script Problem
Month 2 introduced the first complication. The closer team flagged that 28–31% of AI-qualified "high-scored" leads were misqualified — buyers who had confirmed a budget of ₹1.4–1.7 crore but were expecting 3BHK configurations, which the Sector 106 project didn't offer (only 2BHK and 4BHK). The qualification script was confirming budget without confirming configuration compatibility.
The fix: Added a configuration-matching step to the qualification framework. After confirming budget, the AI now cross-references the confirmed BHK preference against available configurations for the specific project and flags configuration mismatches as disqualifying before a high score is assigned.
This revision dropped qualified lead volume temporarily but improved accuracy — misqualification rate dropped from 31% to 11% by end of month 2.
| Metric | Month 1 | Month 2 | Change |
|---|---|---|---|
| Contact rate | 79% | 82% | +3 pp |
| Qualified leads | 89 | 74 | −17% |
| Qualified lead accuracy | 69% | 89% | +20 pp |
| Site visits | 29 | 27 | −7% |
| Bookings | 5.2 | 5.8 | +12% |
Fewer qualified leads, more bookings. The accuracy improvement meant the 27 site visits in month 2 were higher-quality buyers than the 29 in month 1. Site visit to booking conversion improved from 17.9% to 21.5%.
Month 3: The After-Hours Discovery
Month 3 produced the finding that the principal described as "the moment I understood what we'd been leaving on the table."
The operations report revealed that 34% of the brokerage's leads (144 of ~420/month) were arriving between 8 PM and 9 AM. Of these, the human team had been contacting approximately 22% — leads who happened to be in the queue when the 9 AM shift started. The remaining 78% (112 leads) received their first contact attempt more than 12 hours after inquiry submission.
With AI calling extended to the full 6 AM–10 PM window, the after-hours contact rate for those 144 nightly leads rose from 22% to 87%. The 65 additional contacted leads generated 22 additional qualified leads — 7 of which converted to site visits in month 3.
| Metric | Month 2 | Month 3 | Change |
|---|---|---|---|
| Contact rate (overall) | 82% | 88% | +6 pp |
| After-hours contact rate | 22% | 87% | +65 pp |
| Qualified leads | 74 | 96 | +30% |
| Site visits | 27 | 38 | +41% |
| Bookings | 5.8 | 7.1 | +22% |
| Revenue | ₹18,56,000 | ₹22,72,000 | +22% |
Month 3 was the inflection point. The after-hours extension alone accounted for the majority of the month-on-month improvement.
Months 4–6: Stabilisation and Optimisation
Months 4–6 involved incremental refinements rather than step changes.
Month 4 — Dormant lead re-engagement. Activated the re-engagement sequence for leads more than 30 days old. 847 dormant leads from the prior 3 months re-entered the calling queue. Recovered 31 qualified leads from this cohort, generating 9 site visits and 2 bookings — ₹6,40,000 in revenue from leads that would previously have been written off.
Month 5 — Site visit confirmation sequence. Implemented WhatsApp confirmation messages 48h, 24h, and morning-of. No-show rate dropped from 22% to 9%. The 38–42 monthly site visits now produced 34–38 actual attended visits versus 30–33 previously.
Month 6 — Team restructure. Reduced calling team from 4 BDRs + 1 team lead to 2 BDRs + 0.5 FTE operations specialist. The 2 retained BDRs focus exclusively on warm lead management — buyers who are qualified but need multiple touchpoints before agreeing to a site visit. Initial cold calling and qualification is fully AI-operated.
6-Month Results Summary
Pre-Deployment vs Month 6
| Metric | Pre-Deployment (3mo avg.) | Month 6 | Improvement |
|---|---|---|---|
| Contact rate | 44% | 89% | +45 pp |
| Qualified leads/month | 48 | 102 | +113% |
| Site visits/month | 17 | 44 | +159% |
| Bookings/month | 3.1 | 8.6 | +177% |
| Revenue/month | ₹9,92,000 | ₹27,52,000 | +177% |
| Calling cost/month | ₹2,85,000 | ₹92,000 | −68% |
| Total cost (calling + marketing) | ₹4,70,000 | ₹2,77,000 | −41% |
| CAC | ₹1,51,613 | ₹32,209 | −78.8% |
| Net monthly contribution | ₹5,22,000 | ₹24,75,000 | +374% |
What Made the Difference: The Principal's Assessment
In a review at month 6, the brokerage principal identified three factors that drove the outcome.
1. After-hours lead recovery was the biggest single lever. "We thought our 44% contact rate was our contact rate problem. It wasn't — the real problem was 34% of our leads arriving at night when no one was calling. Once AI covered those hours, the contact rate went up and the revenue followed immediately."
2. Qualification accuracy improvement was slow but commercially significant. "Month 2 felt like a step backward when qualified lead volume dropped. But the bookings didn't drop — they went up. That's when I understood that our human team was generating high-quantity, low-quality qualified leads. The AI produced fewer but more accurate qualified leads that actually converted."
3. Re-engaging dormant leads was free money. "We had 847 leads in the CRM older than 30 days that nobody was calling. Activating the re-engagement sequence on those leads in month 4 produced 2 bookings from leads we had written off. That's ₹6.4 lakh from people already in our CRM."
Disclaimer: Financial metrics, operational benchmarks, and month-by-month performance data in this case study are based on an actual Gurgaon residential brokerage deployment through mid-2026. The brokerage's identity has been anonymised at their request. All numerical data is drawn from the brokerage's Sell.Do CRM reporting and monthly operational reviews. Results are specific to this brokerage's lead quality, project portfolio, market position, and team configuration — other brokerages will experience different outcomes based on their own operational variables. This case study does not constitute a performance guarantee for any specific deployment.
Frequently Asked Questions
Related Reading
Written by
Zappio Team
AI & Real Estate Experts — building AI voice agents that qualify real-estate leads in minutes, not days.
See more Zappio insights in Google Search — add us as a preferred source.
Add as preferred source