Try the demo Book a call

Examples · seven conversations, annotated

What a real estate chatbot example looks like.

A real estate chatbot example shows the assistant handling an enquiry end to end: answering from live listing data, qualifying the buyer, booking a viewing, and refusing to steer on protected classes. The seven transcripts below are illustrative demonstrations, not logs of real customers.

Try the live demo Book a 15-min call
7end-to-end transcripts, annotated
Live listingsanswers, not “an agent will call”
Fair Housingrefuses to steer — shown in action
Honestdefers when it doesn’t know

An honest note before you read

These are illustrative examples, not real customer chats. The service is new and has no client conversations to publish, so every buyer, address and price below is made up. What’s genuine is how the assistant is built to behave — the qualifying, the booking, the Fair Housing refusal and the honest defers are exactly how it responds. Want the real thing? The live demo answers whatever you type.

Seven scenarios

The conversations, start to finish.

Each transcript is a full exchange, followed by a short note on what the bot did well. Together they cover the situations that decide whether a chatbot is worth having: after hours, unqualified browsers, sellers, the compliance traps, and the moments it should say “I don’t know”.

RE
214 Maple Court11:48pm · website chat, opened from a listing page
Example
Hi — you’re looking at 214 Maple Court. I’m the assistant here; I can answer questions about it or set up a viewing, even this late. What would you like to know?
Is it still available? And does it have a garage?
It is — it came on two days ago. Two-car attached garage, three beds, two baths and a finished basement. Here’s the listing:
214 Maple Court
214 Maple Court3 bed · 2 bath · 2-car garage$429,000
Nice. Could I see it this weekend?
Absolutely. The agent, Dana, has Saturday 10:30am or Sunday 1:00pm open. Which suits you better?
Saturday works.
Booked for Saturday 10:30. What’s the best number to send the confirmation and address to?
555-0148 — a text is fine.
Done. You’ll get a text shortly and Dana will meet you at the door. Anything else you’d like to know before Saturday?

Example 1 · After hours · answered from the listing, then booked

Answered from the actual listing — then booked a viewing.

Half of all property enquiries arrive after hours, and 78% of buyers work with the first agent to respond. Here the bot did the responding at midnight: it pulled real facts from the listing feed — availability, the garage, bed and bath counts — instead of promising that “an agent will be in touch”, then turned interest into a calendar booking with consent-based contact details.

What it did well

  • Answered from live listing data, not a canned holding reply
  • Booked the viewing into the agent’s calendar
  • Captured the number with explicit consent to text
RE
Facebook MessengerSaturday afternoon · Facebook Messenger
Example
Do you have anything cheaper than the ones on your page?
Possibly — let me point you the right way. Which area are you focused on?
Somewhere around Elmwood, ideally.
Good — we cover Elmwood. What price range should I stick to so I only send ones that fit?
Under 300 really.
That’s workable there. How soon are you hoping to move — weeks, a few months, or further out?
Honestly, maybe next year. Just seeing what’s out there for now.
Totally fair — no rush at all. Have you spoken to a lender yet, or is the financing still to sort out?
Not yet.
No problem. I’ll send a few Elmwood options under $300k and keep you posted as new ones come up — no pressure, and you can tell me to stop any time. Best email for those?

Example 2 · Browser, not a buyer · qualified and routed to nurture

Recognised a next-year browser — and didn’t page an agent.

Not every enquiry deserves an interruption. The bot established area, budget, timeline and financing in four plain questions, realised this was a browser twelve months out, and dropped them into a nurture sequence in the CRM instead of firing an alert at an agent’s weekend. The buyer still gets looked after; nobody’s time is wasted on false urgency.

What it did well

  • Qualified area, budget, timeline and financing conversationally
  • Sorted “someday” from “now” and skipped the agent alert
  • Offered an opt-out — a nod to TCPA-friendly consent
RE
Investment sellerWeekday evening · website chat
Example
I’ve got a rental on Cedar Street I might offload. What’s it worth?
Happy to help you scope that. Is this an investment property you’re selling, or your own home?
Investment. The tenant’s moving out in spring.
Got it. A proper number needs the address and a couple of details — beds, baths and rough condition — so an agent can run a comparative market analysis rather than me guessing. Shall I set that up?
What would I actually walk away with after fees?
That depends on the sale price, your mortgage payoff and closing costs — the agent can model your net proceeds properly, which is worth more to you than a guess from me. Would a 15-minute valuation call work? What days suit?
Thursday after 5.
Thursday evening it is. What’s the best number for the agent to confirm the time and address?

Example 3 · Motivated seller · switched to a valuation appointment

Spotted a seller, switched flows, booked a listing appointment.

The same assistant that qualifies buyers detects sellers too. It recognised an investor offloading a rental, moved off the buyer script into a valuation flow, and — crucially — refused to invent a home value or a net-proceeds figure it couldn’t stand behind. Instead it booked a comparative-market-analysis call, which is where a real listing conversation actually starts.

What it did well

  • Detected a seller and switched from the buyer flow
  • Declined to fabricate a valuation or net-proceeds number
  • Booked a CMA appointment for the agent
RE
214 Maple CourtFrom a listing page · website chat
Example
Is this a good neighbourhood for a family like mine? What’s the demographic around here?
I can’t speak to who lives in an area — steering on things like that isn’t something a fair, compliant assistant should do, and it wouldn’t be a reliable answer anyway. I can give you the objective facts about the location, though.
Okay — like what?
For this address I can share the assigned school ratings, nearby parks and amenities, walk and transit scores, distance to the highway, and recent sale prices on the street. Want me to pull those together?
Yes — schools and prices.
Here are the assigned elementary, middle and high schools with their published ratings, plus the last four sales on the block. If it helps, the agent can walk you through the area’s amenities on a call.

Example 4 · Fair Housing · refused to steer, redirected to facts

The compliance moment — a deliberate non-answer, by design.

This is the transcript that matters most. A demographic or “good for families like mine” question is a Fair Housing trap: answering it is steering, and steering is unlawful. The bot doesn’t rely on tact or luck to dodge it — it’s built with a hard guardrail that refuses the framing and redirects to objective, published facts anyone can verify. A helpful-sounding wrong answer here is a liability; the bot is designed to give the safe one instead.

What it did well

  • Refused a protected-class / steering question outright
  • Redirected to objective facts: schools, transit, sale prices
  • A built-in constraint, not a one-off lucky reply
RE
Facebook MessengerMidday · Facebook Messenger
Example
Can you get me into 88 Birch Lane on Sunday?
I can look into it. Quick check first — are you working with an agent already, or would this be with us?
I’ve got an agent, but she’s away this week.
Ah — in that case the right move is still to go through her; she’ll usually have cover, or can set the viewing for when she’s back. I don’t want to step on an arrangement you already have. Is there anything general I can answer about the home in the meantime?
Fair enough. Just tell me the square footage.
It’s listed at 1,940 sq ft — three beds, two and a half baths. Your agent can arrange the showing; hope it goes well.

Example 5 · Already represented · stepped back politely

Detected another agent — and didn’t poach the lead.

One question — “are you already working with an agent?” — is ethics and efficiency at once. Poaching a committed client is a bad look and a waste of everyone’s time. The bot caught the existing relationship, declined to book around it, and stayed helpful with public information only. It protects your reputation as much as your pipeline.

What it did well

  • Detected an existing agency relationship early
  • Declined to book around another agent’s client
  • Still helped with public listing facts
RE
17 Willow BendLate morning · website chat
Example
Does 17 Willow Bend let me add a second kitchen for an in-law suite? And is the seller open to a rent-back?
Both are good questions I shouldn’t guess at — a second-kitchen conversion comes down to local zoning and permits, and a rent-back is entirely the seller’s call. I’d rather not give you a wrong answer on either.
So who can tell me?
The agent can check the zoning and ask the seller directly. Want me to have them follow up? If you leave the best email or number and a good time, I’ll pass this exact question along.
Sure — mornings are best. jordan@example.com.
Thanks, Jordan. I’ve logged the zoning and rent-back questions for 17 Willow Bend and flagged them for a morning follow-up. You’ll hear back with real answers, not a guess.

Example 6 · Out of its depth · honest defer and human capture

Said “I don’t know” — then captured it for a human.

A confident wrong answer about zoning or a seller’s intentions is worse than no answer. The bot recognised the edge of what it should speak to, refused to improvise, and did the valuable thing instead: it captured the exact question with contact details and handed it to a person. An honest defer keeps trust intact and still moves the lead forward.

What it did well

  • Refused to invent zoning or seller-intent facts
  • Captured the specific question, not just a name
  • Handed a warm, contactable lead to a human
RE
WhatsAppAfter hours · WhatsApp
Example
We’re relocating from Denver — do you cover Boulder County?
Congratulations on the move. I should be straight with you: Boulder isn’t in the area this team serves, so I’d be doing you a disservice pretending otherwise.
Ah, okay.
What I can do is take a few details and have the agent refer you to someone trusted who knows Boulder well — so you’re not starting cold in a new city. Want me to set that up?
That’d be great, thanks.
Nice one. Best email for the introduction, and roughly when you’re moving?

Example 7 · Out of area · honest referral instead of a chase

Flagged an out-of-area lead — and made it useful anyway.

An enquiry the team can’t serve is a referral, not a lead to chase. Rather than fake coverage to keep the conversation going, the bot said so plainly and converted the moment into a referral — respecting the buyer’s time and the agent’s reputation. Honesty here is the premium move, not a missed opportunity.

What it did well

  • Admitted the enquiry was outside the service area
  • Turned a dead end into a referral capture
  • Protected the buyer’s experience and the agent’s name

Transcripts are tidied for reading — real conversations wander, backtrack and mistype more than these do. The wording, service area and qualifying questions are tuned to your market on the discovery call; the flow shown here is the shape, not a rigid script.

Why these six behaviours

Every example is proving one thing.

A chatbot is easy to demo and hard to trust. These transcripts are chosen to show the behaviours that separate a real estate assistant from a scripted FAQ widget — the useful ones and the disciplined ones.

It knows the listings

Availability, garage, beds, price — pulled from your IDX/MLS feed and shown in the chat, so a midnight enquiry gets a real answer instead of “an agent will call”.

It sorts the lead

Area, budget, timeline, financing and agent status — asked conversationally, then routed. Ready buyers get booked; browsers get nurtured; nobody’s paged at the weekend for nothing.

It stays compliant

Demographic and steering questions get a deliberate non-answer and a redirect to objective facts. Fair Housing isn’t left to tact — it’s a built-in guardrail.

It knows its lane

Already have an agent? It steps back. Out of area? It refers. It won’t poach a committed client or chase an enquiry the team can’t serve.

It defers honestly

Zoning, permits, a seller’s intentions — when it shouldn’t guess, it says so and captures the exact question for a human, rather than risking a confident wrong answer.

It closes the loop

Every path ends in something concrete — a booked viewing, a valuation call, a nurture sequence or a flagged follow-up — handed to your calendar and CRM.

What these examples don’t show

Where a curated transcript stops — and the bot defers.

A tidy example is a best case. Knowing what it leaves out is the honest way to judge whether this fits your business.

  • Real chats are messier. Live buyers go off-script, mistype, change their minds and ask three things at once. These transcripts are cleaned up for reading; the demo is closer to the real, untidy thing.
  • It doesn’t negotiate or advise. No pricing strategy, offer tactics, or legal and financial guidance — as the seller and zoning examples show, that’s the agent’s work and the bot hands it over.
  • It only knows what it’s connected to. Answers are as current as your IDX/MLS feed and the facts you give it. It can’t see inside a home, verify a self-reported budget, or know a seller’s private intentions.
  • Guardrails beat helpfulness. Some questions get a deliberate non-answer — Fair Housing and TCPA constraints mean it won’t improvise past its limits, even when a “helpful” reply would be easy to fake.
  • No numbers to show yet. These illustrate handling, not conversion rates or income. The business is new; we’d rather show you exactly how it behaves than invent statistics we can’t stand behind.

About these examples

Straight answers on the transcripts.

The questions people reasonably ask when a page shows them conversations. If yours isn’t here, ask it on the call.

Are these real customer conversations?

No — and we won’t pretend otherwise. These are illustrative, representative examples written to show how the assistant is built to respond in common situations. The service is new and has no client chats to publish, so the buyers, addresses and prices are made up. What’s real is the behaviour: answering from live listings, qualifying budget, area and financing, refusing to steer on protected classes, and deferring honestly when it doesn’t know. If you’d like to see it for yourself, the live demo responds to whatever you type.

Can I see the chatbot live instead of reading transcripts?

Yes. The live demo lets you play the buyer and watch the bot answer, qualify and offer a booking in real time. Reading a transcript shows you the shape of a conversation; the demo lets you try to break it — ask an off-topic question, a demographic question, or a listing detail, and see how it holds up.

Why does the bot refuse the “good neighbourhood for families” question?

Because answering it would be steering, and steering by protected class breaches the Fair Housing Act. The assistant is deliberately built to decline demographic and “is it right for people like me” questions and redirect to objective, published facts — school ratings, transit scores, amenities, recent sale prices. It’s a hard design constraint, not a judgement call, which is exactly why you can rely on it. The same discipline shows up in how it qualifies leads on objective criteria only.

What happens when the bot doesn’t know the answer?

It tells you it doesn’t know and captures the question for a human, rather than guessing. Zoning, permits, a seller’s intentions, financial or legal specifics — the bot recognises these as outside its remit, logs the exact question with the buyer’s contact details, and flags it for an agent to follow up. An honest defer protects trust and still hands you a warm, contactable lead.

Can these examples be customised to my listings and market?

Yes — the wording, service area, qualifying questions and listings are all tuned to you on a short discovery call, and the bot connects to your own IDX/MLS or website feed so it answers with your actual properties. Plans start at Solo Assistant ($750 build + $250/mo) and scale to Lead Engine ($1,500 + $350/mo) and Brokerage ($2,500+ + $500/mo); see pricing for what each includes.

See it handle your buyers, not ours

The transcripts show the shape; the demo lets you test it. Play the buyer, or book a 15-minute call and we’ll run these scenarios against your own listings.

Try the live demo Book a 15-min call

Keep exploring

Try the live demoPlay the buyer and watch it answer, qualify and book in real time. Lead qualificationThe seven-question script behind how it sorts every enquiry. PricingSolo Assistant, Lead Engine and Brokerage — what each plan includes.
Chat on WhatsApp