AI Receptionist for Realtors: What It Books, What It Costs, Where It Fails
The short answer
An AI receptionist for realtors answers every call to your listings, qualifies the caller, books the viewing into your calendar, and follows up by text — 24 hours a day, including the evenings and weekends when most buyers actually call. For a solo agent or a small team, it costs roughly $50–$450 a month depending on call volume, which is a fraction of one lost commission. It is not a replacement for you: it cannot negotiate, it cannot read a nervous seller, and it must never be allowed to answer a fair-housing question with a made-up answer. Used as a first responder that books and qualifies, it fixes the single most expensive leak in a real estate business — the call nobody answered.
That leak is bigger than most agents think. The National Association of Realtors (NAR) reports that 78% of homebuyers end up working with the first agent who responds to their inquiry, and around 70% of buyers interview only one agent before choosing. The agent who picks up first usually wins the client. The problem is that the average agent takes about 15 hours to respond to a new lead — and nearly half never follow up at all. An AI receptionist exists to close that gap, not to replace the agent on the other side of it.
The missed-call problem, with honest numbers
Real estate is a speed business built on a phone that rings when you are least available: during a viewing, in the car between appointments, at dinner, on a Sunday afternoon when a buyer drives past a listing and calls the number on the board. Every one of those calls is a buyer or seller who has already decided they are interested. The numbers on what happens next are uncomfortable, and they come from different kinds of studies, so read them with their sources attached.
- 78% of buyers work with the first agent who responds. NAR’s Profile of Home Buyers and Sellers. Around 70% of buyers interview only one agent. First response is not a nice-to-have; in most transactions it is the whole contest.
- The average agent takes 917 minutes — over 15 hours — to respond to a new lead, and 48% never follow up at all. Inman / AgentZap 2025–2026 industry survey data. This is survey data from a vendor-adjacent source, not NAR’s own measurement, but it points the same way as every other study in this section.
- 41% of top brokerages never responded at all. Roof AI secret-shopped 74 of the largest US brokerages: only 9% responded within five minutes, and 41% never answered the inquiry. These were the big, well-resourced firms.
- Contacting a lead within five minutes makes you 21 times more likely to qualify it than waiting 30 minutes, and the odds of ever reaching the lead fall roughly 100-fold after half an hour. MIT / InsideSales.com Lead Response Management study, based on 15,000+ web leads and 100,000+ call attempts. The same pattern shows up in the Leads360/Velocify data: leads called within 60–120 seconds converted 391% above average.
- 23% of companies never answered a test web lead at all, and the average reply took 42 hours. Harvard Business Review, 2011 audit of 2,241 US companies. Old, and not real-estate-specific — but it is the study every speed-to-lead claim traces back to, and nothing since has contradicted it.
- 44–62% of real estate inquiries arrive outside business hours. Estimates vary by source — NAR/Zillow-cited figures put it above 60%, other industry surveys around 44% — but every version agrees on the direction: a large share of your leads arrive when your office is closed.
Put those together and the shape of the problem is clear. A buyer calls three listings in a row on a Saturday evening. The first agent who answers gets the viewing. The other two get a voicemail nobody returns, because by Monday the buyer has already toured with agent number one. The 917-minute average response time is not a staffing problem you can fix by trying harder — you are one person, and you are with clients during the hours buyers call. It is a coverage problem, and coverage is exactly what an AI receptionist sells.
One more number matters for the money math later: a single transaction side is worth real money. With a median US home price around $400,000 (NAR, 2025) and a typical buy-side or list-side commission near 2.5–3%, one converted lead is worth roughly $10,000–$12,000 in gross commission before your brokerage split. A missed call is not a minor inconvenience. At portal lead prices of $200–$450 each (Zillow and Realtor.com leads, per industry reporting), letting a $300 lead die in voicemail is the most expensive habit in the business.
What an AI receptionist for realtors actually books
A good AI receptionist for a real estate business does four jobs, and you should judge every vendor on these four specifically — not on how human the voice sounds in a demo.
1. Viewings and open-house bookings
This is the core job. The caller asks to see a property; the receptionist checks your calendar, offers the open slots, books the viewing, sends a confirmation text with the address and your name, and adds a reminder the morning of the viewing. It should also handle the follow-ups that surround a viewing: rescheduling, cancelling, asking for directions, and confirming the buyer is still coming. If the vendor cannot show you a live calendar integration with Google Calendar or Outlook, it is not a booking tool — it is a voicemail with a nicer voice.
2. Buyer qualification
Before a viewing gets booked, the receptionist asks the questions you would ask: are you pre-approved or pre-qualified, what is your budget range, what is your timeframe, are you working with another agent? The answers go into your CRM as structured notes, not as a transcript you have to read. The point is not to interrogate the caller — it is to stop you spending Saturday driving to a viewing for a buyer who cannot get a mortgage and has not spoken to a lender.
There is a hard line here, and it is the law. The receptionist must never steer a caller toward or away from a neighbourhood on the basis of race, religion, family status, disability, or any other protected characteristic. That is a Fair Housing Act violation whether a human or a machine says it. Any vendor you consider must show you, in writing, how their system handles questions about schools, crime, demographics, and “good neighbourhoods” — the correct answer is a neutral, factual redirect to public data sources, never an opinion. If they cannot show you that, walk away.
3. Seller and valuation enquiries
The second-most valuable call a real estate office gets is “what is my home worth?” The receptionist takes the address, the caller’s timeframe for selling, and a few facts about the property, then books a valuation appointment and sends the caller a confirmation. It should also catch the callers who are not ready yet — “we are thinking about next spring” — and put them on a nurture follow-up instead of letting them drift. NAR’s 2025 data shows 91% of sellers use an agent, and most interview only one. The valuation call is the listing appointment, one step removed.
4. The calls that are not buyers at all
A surprising share of an agent’s phone traffic is vendors, contractors, photographers, mortgage brokers, and other agents. A receptionist that routes these — takes a message with the right details, books the photographer, forwards the urgent contractor call — saves the interruptions that eat a working day. It should also screen the obvious spam and robocalls, which are a real tax on a publicly listed number. And it should handle rental enquiries separately, with their own script, if your office does lettings — a renter asking about deposits and pet policies is a different conversation from a buyer asking about a viewing.
What it costs in 2026
Pricing for AI receptionists follows the same patterns as the rest of the market: per-minute plans, per-call plans, and flat-rate unlimited plans. The table below uses vendor pricing pages checked in September–November 2026, and the same vendors serve home services and dental practices, so the prices are real — but check them before you buy, because AI voice pricing moves every quarter.
| Vendor | Typical plan | What you get |
|---|---|---|
| Rosie | $49/month for 250 minutes | Budget entry point; fine for a solo agent with light call volume. |
| Goodcall | $79/month for 100 callers | Per-caller pricing; watch the definition of a “caller” before you sign. |
| Smith.ai | $150/month for 75 calls (AI tier) | Hybrid option: AI answers, humans available as backup. |
| NextPhone | $199/month flat | Flat-rate unlimited; predictable if your volume is high. |
| Sameday AI | $449/month for 500 minutes | Higher tier with deeper CRM and scheduling integrations. |
| Jobber Receptionist | $29/month add-on | Only makes sense if you already run Jobber; limited to its ecosystem. |
Most solo agents land in the $79–$199 range. Most small teams land in the $199–$449 range. Compare that with the alternatives: a part-time human receptionist costs $1,500–$2,500 a month before payroll taxes, a live answering service runs $200–$400 a month for a small call volume and still cannot book into your calendar properly, and your own time — the hours you spend playing phone tag — is the most expensive option of all.
Two pricing traps to watch. First, per-minute plans punish long calls, and real estate calls run long: a buyer asking detailed questions about a property can take eight minutes. Second, some vendors charge extra for the integrations that make the product useful — CRM sync, calendar booking, and texting are sometimes add-ons, not included. Get the all-in monthly number in writing, including overage rates, before you compare vendors.
The ROI math for a solo agent
Worked example, with the assumptions stated so you can redo it with your own numbers. Take a solo agent in a mid-priced US market. Median home price $400,000, commission 2.5% per side, so one closed side is worth $10,000 gross before the brokerage split — call it $7,000 net to the agent after a 70/30 split.
Suppose the agent misses 20 calls a month — a conservative number for an agent with listings live, given that NAR’s own data says most inquiries arrive outside office hours. Of those 20, say half are genuine buyers or sellers rather than vendors and spam. Of those 10, say three would have booked a viewing or valuation. Of those three, say one closes within the year — a low conversion rate for a warm, phone-qualified lead, but let us be pessimistic.
One recovered closing a year is $7,000 net. The receptionist costs $150 a month, or $1,800 a year. That is roughly a 4x return on a deliberately pessimistic model. At two recovered closings a year — still modest — the return is about 8x. The break-even point is one recovered closing every four years, which is an absurdly low bar: if the system books a single viewing that becomes a single sale in four years, it has paid for itself.
The honest version of this math cuts both ways. If your phone barely rings — a new agent with no listings and no marketing — a receptionist will not create demand. It captures demand that already exists. The agents who get the best return are the ones with listings live, portal leads running, and a phone number on a sign board. If that is not you yet, spend the $150 on lead generation first and come back to this when the phone starts ringing.
Where an AI receptionist for realtors fails
The failures are specific to real estate, and you need to hear them before the demo, not after the contract.
Negotiation and distressed sellers. A seller calling because their sale fell through, a buyer who just lost a bidding war, an executor selling a parent’s house — these callers need a human who can hear distress and respond with judgement. An AI that cheerfully offers to book a viewing will damage the relationship. Your receptionist needs an escalation rule: emotional language, words like “fell through”, “divorce”, “probate”, or “urgent” should route to you immediately, not into a booking flow.
Fair housing and steering. Already covered above, but it is the failure with legal consequences. The system must never describe a neighbourhood’s character, schools, or residents in subjective terms. Test this yourself in the trial: call in and ask “is this a good area for families?” and “what kind of people live there?” If the answer is anything other than a neutral redirect to public data, do not buy it.
Complex or high-value questions. “What are the HOA special assessments on this building?” “Is the roof under warranty?” “Can I see the seller’s disclosure?” The receptionist does not know, and a confident-sounding wrong answer is worse than no answer. The right behaviour is to take the question, promise a callback with the answer, and put it in front of you with the caller’s details. Any vendor that claims their AI can answer property-specific questions from your listings without a live data feed is overselling.
Lockboxes, access, and the physical world. It cannot let anyone into a property, reset a lockbox code, or confirm that the keys are where they should be. If your workflow depends on the person answering the phone also handling access, the receptionist needs a clear handoff: who gets called, in what order, when access is the question.
Accents, names, and addresses. Voice AI still struggles with unusual street names, non-Anglo names, and strong accents — exactly the callers you least want to frustrate. Ask the vendor for their accuracy on address capture specifically, and test it with your own hardest street names during the trial.
Compliance and consent. Follow-up texts and calls fall under TCPA rules in the US. Your receptionist should only text callers who have given consent or are responding to their own enquiry, and it should honour opt-outs instantly. If the vendor cannot explain their TCPA handling, that is a red flag, not a detail.
AI receptionist vs live answering service vs hiring
| AI receptionist | Live answering service | Part-time receptionist | |
|---|---|---|---|
| Monthly cost | $50–$450 | $200–$400 | $1,500–$2,500+ |
| Available | 24/7 | 24/7 (often) | Office hours only |
| Books viewings into your calendar | Yes, live | Usually takes a message | Yes |
| Qualifies buyers (budget, pre-approval) | Yes, scripted | Rarely | Yes, with training |
| Handles distressed or complex callers | Poorly — needs escalation rules | Moderately | Well |
| CRM notes | Automatic, structured | Message transcription | Manual |
| Scales with call volume | Instantly | Yes, at a price | No — one person |
The honest read: a live answering service is a polite message-taker, and a part-time hire is the best experience at five times the price. The AI receptionist wins on booking and availability per dollar, and loses on judgement. For a solo agent, the practical answer is usually the AI receptionist plus a clear escalation rule to your own phone — you get the coverage without paying for a person to sit by a phone that rings ten times a day.
The 7-question checklist before you buy
- Does it book directly into my calendar, live, during the call? Not “sends me a notification to book it myself”. Live, or it is not worth buying.
- Does it write structured notes into my CRM? Name, number, property, budget, pre-approval status, timeframe — as fields, not a transcript. If it cannot integrate with your CRM (Follow Up Boss, kvCORE, Salesforce, HubSpot, or even a spreadsheet via Zapier), you will be retyping everything.
- Show me the fair-housing handling, in writing. How does it answer “is this a safe neighbourhood?” and “what kind of people live here?” If the vendor cannot show you, do not buy.
- What is the escalation rule? Which words, tones, or topics route the call to you immediately, and how fast does that handoff happen? Test it with a distressed-seller scenario in the trial.
- What is the all-in monthly cost at my volume? Base plan, overage rates, integration fees, texting fees. Get it in writing. Then check it against your missed-call volume, not the vendor’s estimate of it.
- How does it handle addresses, names, and accents? Test it with your hardest street names. Ask for accuracy figures on address capture, not a demo with a perfect American accent.
- What happens when it fails? Every system fails sometimes. Does the caller get a graceful fallback to voicemail or to you, or do they get a dead line? Ask to hear the failure path, not just the happy path.
How to roll it out without annoying your customers
Start with after-hours coverage only. Let the receptionist answer when you cannot — evenings, weekends, and while you are in appointments — and keep answering the phone yourself during office hours for the first two weeks. That way you hear the receptionist’s work in the morning (every call logged, every booking in your calendar) before any of your regular clients meet it.
Then tell your callers. A short line on your voicemail and your email signature — “If I’m with a client, my assistant will take your details and book you in” — sets the expectation. Nobody minds an AI answering the phone when it books them a viewing in thirty seconds. People mind it when it pretends to be you, wastes their time, or traps them in a loop. Configure the greeting to be honest about what it is, keep the booking flow under a minute, and always offer a callback from you as an option.
Review the first fifty calls yourself. Listen to the recordings, read the notes, and fix the script: the questions it asks in the wrong order, the property names it mishears, the callers it should have escalated. Two weeks of that and the system will be better than any new hire’s first month — because unlike a hire, it applies the fix to every call from then on.
Frequently asked questions
Will buyers know they are talking to an AI?
Probably, and that is fine. Most callers can tell within a few seconds, and the research on this is consistent: people do not mind an AI that solves their problem quickly. What they mind is an AI that pretends to be human, wastes their time, or cannot get them to a person. Be honest in the greeting, keep the booking fast, and always offer a callback.
Can an AI receptionist replace my ISA or inside sales agent?
No. An inside sales agent nurtures leads over weeks, handles objections, and builds a relationship. A receptionist handles the first sixty seconds: answer, qualify, book. If you already have an ISA, the receptionist makes them more effective by handing them qualified, booked appointments instead of cold lead lists. If you do not have one, the receptionist is the cheaper first step — it captures the lead, and you do the nurturing.
Does it work with Zillow, Realtor.com, and portal leads?
It works with the phone calls those leads generate, which is where the money is. A portal lead that calls your number goes straight to the receptionist, which qualifies and books them before they call the next listing. The portal lead itself still arrives in your CRM by email; the receptionist does not replace your lead source, it answers the phone when the lead decides to call instead of filling in a form.
What about fair housing — is an AI receptionist a legal risk?
It is a risk only if it is configured badly. The law applies to the brokerage, not the software vendor, so the responsibility is yours. The protection is a tested script: neutral, factual answers to neighbourhood questions, immediate escalation of anything that touches a protected characteristic, and a regular review of call recordings. Ask the vendor for their fair-housing documentation before you sign, and test the edge cases yourself in the trial.
How much does an AI receptionist for realtors cost?
Most solo agents pay $79–$199 a month; small teams pay $199–$449. Entry plans start around $49 a month for low call volumes. One recovered closing — roughly $7,000–$10,000 net on a median-priced home — covers several years of the service. The break-even is so low that the real question is not the price, it is whether your phone rings enough to justify it.
Can it handle rental enquiries as well as sales?
Yes, with a separate script. Rental callers ask different questions — deposits, pet policies, lease terms, move-in dates — and they should get different answers and a different booking flow. If your office does both sales and lettings, make sure the receptionist can tell the two apart and route them correctly. The vendors in the pricing table above all support multiple scripts; confirm it is included in the plan you are buying, not an add-on.
The bottom line
Real estate rewards the agent who answers first, and punishes the one who answers on Monday. NAR’s data says 78% of buyers work with the first agent who responds; the industry’s own surveys say the average response takes fifteen hours. That gap is the whole business case for an AI receptionist for realtors. It is not a robot agent. It is a first responder: it answers, it qualifies, it books the viewing, and it gets out of the way. The judgement — the negotiation, the distressed seller, the deal itself — is still yours, and it always will be.
If your phone rings and nobody answers it, that is the leak to fix first, before the next marketing spend. If you want help setting one up — choosing a vendor, writing the script, wiring it into your calendar and CRM — talk to us. We build AI receptionists and voice agents for small businesses, and we would rather tell you the honest answer about whether one will pay for itself than sell you one that will not.
Related reading: How Much Does an AI Receptionist Cost in 2026? · Missed Calls Are Sending Customers to Your Competitor · AI Receptionist for Dental Practices · AI Receptionist for Home Services