Replying to Property Leads in Under a Minute: How a Real Estate Agency Turned Website Chats into 310 Booked Viewings a Month
Oct 30, 2026
4 min read
Oct 30, 20264 min readBusiness context
This client is a residential real estate brokerage in the US Southeast with 140 agents across four offices, handling buyer, seller, and rental business. Its website, which carries an IDX listing feed, generates about 1,900 chat and form inquiries a month, roughly 42% of them in the evening or on weekends. Two inside sales coordinators took the first touch on every inquiry, then passed qualified leads to agents by area. During business hours, the first reply took more than three hours on average; inquiries that came in at night waited until the next morning, by which time many buyers had already booked a showing with another brokerage. Only about 5% of website inquiries became a booked viewing. The brokerage needed every inquiry answered while the buyer was still on the listing, qualified consistently, and booked onto the right agent's calendar, while staying fully compliant with fair housing rules.
- Residential Real Estate
- USA
- 140 agents
- 4 offices
- ~1,900 web inquiries/month
- Lead qualification
- Viewing booking
- Conversational AI agent
Challenges
Slow first response
The average first reply during business hours took 3 hours 40 minutes, and evening and weekend inquiries waited 12 hours or more.
Leads that never heard back
About 31% of website inquiries received no reply within 24 hours during busy weeks, mostly the ones sent after the coordinators had logged off.
Inconsistent qualification
Whether a lead was asked about budget, pre-approval, or timeline depended on who picked it up, so agents often drove to showings with buyers who were months away from purchasing.
Scheduling by phone tag
Booking a viewing took an average of four back-and-forth messages between the coordinator, the buyer, and the listing agent.
Missed showings
About 24% of booked viewings ended as no-shows because no one confirmed the appointment on the day.
Solutions we implemented
We built the brokerage's lead response workflow in CogniAgent Applications and deployed it on the website Web Widget and SMS through Twilio, connected to the brokerage CRM, the IDX listing feed, and agents' Google Calendars. The rollout ran in three phases between March and May 2026: discovery with the sales manager and coordinators, configuration and a fair housing review of every question and response, and go-live with coordinators monitoring conversations for the first three weeks. Agents keep every relationship; the agent handles the first touch, qualification, and booking.
An answer while the buyer is still on the listing.
A Channel Dispatcher in responder mode picks up every Web Widget chat and form submission, and the agent replies within seconds with details about the listing the visitor was viewing, pulled from the IDX feed through an HTTP Request.
Natural qualification with the Skillset Engine.
A Communication Actor collects the brokerage's qualification fields as a conversation: buy or rent, target areas, price range, bedrooms, move timeline, financing or pre-approval status, and whether the buyer already works with an agent. It understands informal answers like "somewhere near the good schools, under 500".
Fair housing guardrails built in.
The agent is configured never to ask about or comment on protected characteristics and never to describe neighborhoods by demographics. Questions it cannot answer from approved content go to a licensed agent.
Routing to the right agent.
An LLM node turns the conversation into structured output, and a Multi-Condition routes the lead by area, price band, and lead type to the matching agent pod, with round-robin assignment inside each pod.
Booking on the agent's real calendar.
An Integration Action checks the assigned agent's Google Calendar and offers open slots in the conversation, so the buyer books a viewing without a single call. The lead, transcript, and appointment are written to the CRM.
Confirmations that cut no-shows.
A Pause node holds the workflow until the day of the viewing, then Send Message texts a confirmation with the address and the agent's name, and lets the buyer reschedule in the same thread.
Results
| Before | After | |
|---|---|---|
| First response time | 3 hrs 40 min average, next morning after hours | Under 1 minute, any hour |
| Inquiries unanswered after 24 hrs | ~31% in busy weeks | None |
| Inquiry-to-viewing conversion | ~5% | ~16% |
| Messages to book a viewing | ~4 back-and-forths | Booked in the chat |
| Viewing no-show rate | ~24% | ~11% |
| Booked viewings every month | 95 | 310 |
Replies in under a minute, any hour
Every website inquiry gets a relevant answer while the buyer is still on the listing, including the 42% that arrive in the evening or on weekends.
3.3x more booked viewings
Monthly viewings booked from website inquiries grew from about 95 to 310, with no additional coordinator headcount.
Conversion from 5% to 16%
Fast replies and consistent qualification turned three times as many inquiries into viewings.
Agents meet qualified buyers
Every viewing arrives with budget, timeline, financing status, and the conversation transcript in the CRM, so agents know who they are meeting.
No-shows cut by more than half
Day-of SMS confirmations with easy rescheduling brought the no-show rate from about 24% to 11%.
Coordinators moved from first touch to nurture.
With first response and booking handled, the two inside sales coordinators now focus on long-timeline buyers and seller leads, the conversations that need a person.