Overview
Housing.com is one of India's largest online real estate platforms, running one of the highest-volume buyer funnels in proptech. Every day, thousands of fresh leads come in against an inventory of properties that are not actively being sold. The job is to qualify each lead on location, budget, and BHK, match them to similar live properties, and capture intent before the buyer moves on.
In partnership with SquadStack.ai, Housing.com transitioned this lead qualification motion from human telecalling to a Voice AI agent. The pilot ran for 6 weeks on a focused buyer lead multiplication and has since extended to broker repost calling and dormant seller activation.
The Challenge
Housing.com's lead qualification operation had three constraints that compounded on each other:
Coverage ceiling on humans. Human agents could only attempt each fresh lead, ~1.65 times a day, leaving connectivity stuck at 54-58%.
Narrow intent window. Buyer intent on a fresh listing decays within hours. By the time human agents dialed back, buyers had often gone cold.
CAC pressure. Paid lead volumes were growing faster than headcount could scale.
The Solution
SquadStack deployed a Voice AI agent as the primary qualification layer for fresh buyer leads.
1. The breakthrough: pitch-first conversation design
The first version of the agent profiled buyers in full before recommending a property. We A/B tested a second version that flipped the order: light profiling, then an immediate recommendation, then deeper profiling to refine the match.
The pitch-first version delivered 27% more verified intent. Buyers stayed on the call longer when they heard a real, specific property in the first 30 seconds.
2. A workflow built to compound
Three changes to the calling logic, each closing a different leak.
- 2x attempt intensity: Each lead is attempted 3.3 times on average, twice the human baseline.
- ADC retry agent: A dedicated AI agent retries abruptly disconnected calls (the largest leakage point in any voice campaign), lifting verified intent on the 2nd connected attempt by 67%.
- Warm handoff: At the end of every qualified call, the agent offers the buyer a direct call from the developer of the matched property.
Real customer conversations
The Impact
- Connectivity climbed from 54% to 84% as Voice AI scaled, a 55% lift over the human baseline.
- CAC dropped 30% driven by attempt density and connectivity rather than corners cut on call quality.
- 2x attempt intensity per lead (3.3 vs 1.65 on humans)
- 15.3% lead qualification rate on attempted leads, in line with the human range of 14.4-16.1%
Looking Ahead
With buyer qualification stable, the partnership is scaling to broker repost calling and dormant seller activation.
