Best Prompts for Realtors and Real Estate Agents (2026): Copy-Ready Prompts for Listings, Follow-Up, and Open Houses

The short answer: the best prompts for realtors are not clever tricks. They are short, specific instructions that turn the facts you already have, from an inquiry, a listing appointment, or a showing, into a finished message you can check and send in minutes. Start with the eight prompts on this page, one for each job that decides whether business moves forward: a new buyer lead reply, a seller lead follow-up, a listing description built only from property facts you supply, an open house invitation and reminder, a showing feedback request, a market update to your sphere list, a past client check-in, and a review request. Fill in the brackets, check every fact against your own notes before you send, and save the versions that work in a prompt library such as the tools in this directory's AI Prompt Tools category, so you stop rewriting the same messages from scratch. This page is page two of this prompt library. Page one, for contractors and home services businesses, is at Best Prompts for Contractors and Home Services Businesses (2026).

That speed matters because buyers and sellers rarely wait. A speed to lead guide from Pinova, summarizing the foundational study by Dr. James Oldroyd of MIT with Kristina McElheran and David Elkington in Harvard Business Review in March 2011, reports that the study analysed 1.25 million sales leads and found that firms that attempted contact within five minutes of a web inquiry were 100 times more likely to make two way contact and 21 times more likely to qualify the lead than firms that waited 30 minutes. The same guide reports, from Inman's 2025 Real Estate Technology Survey, that the average real estate agent takes 917 minutes, over 15 hours, to respond to a new lead inquiry. A prompt cannot make the call for you, but it can remove the blank page that slows the reply down.

One rule runs through every prompt on this page, and it is not optional. Fair housing note: describe the property and the facts, never people or protected classes. A fair housing guide from KONAH Media Group puts the federal rule plainly: the Fair Housing Act makes it illegal to advertise housing in a way that shows a preference, limitation, or discrimination based on protected characteristics, which federally include race, color, national origin, religion, sex, familial status, and disability, and state and local rules may protect more. The practical habit the guide recommends is to write about the house, the land, the features, and the location's amenities, and to ask of every phrase whether it describes the property or who you think should buy it. Every prompt below carries that instruction in its own rules, because AI written copy is held to the same standard as copy you write yourself. This page is general information, not legal advice. Follow your brokerage policy and ask your broker when wording is in doubt.

What makes a realtor prompt actually useful?

A useful realtor prompt does four things every time. First, it names the job in plain words, such as drafting a first reply to a buyer who asked about one listing, instead of asking for a generic follow-up message. Second, it carries the facts the model is allowed to use, pasted from your notes, and it tells the model not to invent prices, dates, property features, market figures, or a home's value that are not in those notes. Third, it sets the output shape: a length, a tone, and exactly what the message must include, such as the property they asked about and one clear question. Fourth, it ends with a check step, asking the model to flag anything it had to guess so you can fix it before sending.

Weak prompts fail in predictable ways. They ask for a professional email with no facts, so the model fills the gaps with invented details. They ask for a listing description with no property facts, so the model invents features and reaches for phrases about who should live there, which is exactly the fair housing risk described above. They leave out the area, the timeline, and what the person actually asked, so the output could belong to any agent in any market. Every prompt on this page is built to avoid those failures: one job per prompt, your facts in brackets, a fixed output shape, and a human check before anything reaches a buyer, a seller, or your sphere.

If you want the wider tool picture behind these prompts, start with Best AI Tools for Real Estate Agents in 2026 and AI Tools for Real Estate Marketing (2026), which cover the software categories an agent uses day to day. This page is the companion to that stack: the words themselves, ready to paste into ChatGPT, Claude, Gemini, or the prompt tool you already use.

Which prompt should you use when a new buyer lead comes in?

Use this prompt the moment a portal, website form, text, chat, or referral inquiry lands and you need a first reply that moves the conversation forward instead of just saying thanks. The Pinova speed to lead guide makes the same point this prompt is built on: speed to lead is the time between a prospect submitting an inquiry and the first meaningful outreach, and a generic thanks message with no next step does not count as meaningful. Paste the inquiry details where the brackets are, name the exact property or search they asked about, and send the result within minutes. For the systems that deliver this reply automatically, see AI Tools for Lead Follow Up (2026).

You are helping [agent name], a real estate agent with [brokerage name] in [city and state]. A new buyer lead just came in. Write a short first reply we can send by text and email.

Lead details from our inquiry form:
Name: [lead first name]
What they asked about: [property address or listing, or the area and property type they described, in their words]
What they said they need: [their words, or not stated]
Timeline they stated: [timeline, or not stated]
Source: [portal, website form, referral, open house, past client]

Rules:
1. In the first sentence, name the exact property or search they asked about, in their words.
2. Name the person sending it: [agent first name], from [brokerage name].
3. Offer exactly two times for a call or a showing conversation: [time option 1] and [time option 2]. Do not invent other times.
4. End with one easy question they can answer with a short reply.
5. Keep the text version under 60 words. Keep the email version under 120 words, with a subject line.
6. Do not invent a price, a property feature, a showing time, or any fact that is not in the notes above. If something is missing, leave a bracket for us to fill.
7. Describe property facts only. Never describe people, and never state or imply a preference based on race, color, national origin, religion, sex, familial status, disability, or any other protected class.
8. Tone: friendly, direct, local. No pressure and no hype.

After the two versions, list anything you had to guess, under the heading Check before sending.

The check list at the end is not decoration. On a busy day it is the difference between a fast reply and a fast mistake, because the model will show you the one detail it was unsure about instead of hiding it inside a polished sentence.

What should you say to a seller lead who has gone quiet?

Follow up with something useful, on a schedule, instead of asking again and again whether they got your message. A seller lead is usually deciding on timing, on preparation work, and on who they trust to price and present their home, and a bare check-in answers none of that. This prompt writes a short sequence in one pass, tied to the concern the seller actually stated. It also carries a hard valuation rule: the model must never estimate a home's value, a list price, or a sale price. If a value is needed, the prompt tells the model to use only the data you supply, such as comparable sales you paste in, and to leave a bracket where that data is missing.

You are writing follow-up messages for [agent name] about a seller lead that has gone quiet.

Facts:
Seller: [seller first name]
Property: [property address or area, and property type]
First conversation on: [date], and what they said then: [their words on timing, preparation, or concerns]
Their stated concern or priority: [timing, preparation work, pricing questions, another decision maker, not ready yet, or not stated]
One useful detail we can add now: [a preparation step, a comparable sale we supply below, a calendar fact such as a date they named, or an offer to walk the property]
Comparable sale data we supply, if any: [paste comparable addresses, sale dates, and sale prices exactly as recorded, or write none supplied]
Sender: [agent first name and brokerage]

Write three short messages, labeled Day 3, Day 10, and Day 24:
1. Day 3: confirm you are still available to help on their timeline, and offer a short call to answer the concern they stated.
2. Day 10: lead with the useful detail above, tied to their stated concern.
3. Day 24: a polite closing message that keeps the door open and asks whether their timing changed, with no pressure.

Rules:
1. Each message under 70 words, written for text or email.
2. Never estimate or state a value, a list price, or a likely sale price for their home. Use only the comparable sale data supplied above, exactly as written. If no data was supplied, do not mention value or price at all, and add a bracket that says [NEEDS AGENT INPUT: comparable data].
3. Do not invent property features, market figures, deadlines, or other agents' activity.
4. Describe property facts only. Never describe people or state a preference based on any protected class.
5. Do not repeat the same opening line twice.

Notice what this sequence never does: it never buys a reply with a number. A value given without your own comparable data and your own judgment is not a follow-up, it is a guess with your name on it, and sellers remember the agent whose first number could not be supported.

How do you turn property facts into a listing description?

Start from facts you supply, and only from those facts. The KONAH fair housing guide recommends replacing phrases about people with the features behind them: instead of a phrase about who a home suits, write the bedrooms, the layout, and the yard. The same guide advises stating factual school assignments with a note to verify with the district rather than describing schools in terms of a type of family, and pointing buyers to public sources rather than characterising a neighbourhood yourself. This prompt follows that order. It also refuses to add any feature, room, size, or price that your facts do not state.

You are drafting a listing description for [agent name] at [brokerage name]. Use only the property facts below. Do not add features, rooms, sizes, prices, or terms that are not in the facts.

Property facts we supply:
Address or area as it will appear in the listing: [address or area]
Property type and layout: [type, bedrooms, bathrooms, levels, exactly as recorded]
Size and lot facts: [square footage, lot size, exactly as recorded, or not supplied]
Interior features: [features from our notes, line by line]
Exterior and lot features: [features from our notes, line by line]
Parking and storage: [facts, or not supplied]
Location facts: [named amenities, parks, transit stops, or shops with distances as recorded, or not supplied]
Assigned schools as recorded, if any: [school names exactly as recorded, or not supplied]
Price and terms as recorded, if any: [price and terms exactly as written, or not supplied]
Showing or offer instructions we supply: [instructions, or not supplied]

Write the description in this order:
1. Open with the single strongest property feature from the facts above.
2. Describe the main living areas using only the layout and features supplied.
3. Describe the exterior, lot, parking, and storage using only the facts supplied.
4. Close with one plain location sentence using only the location facts supplied. If schools are supplied, state the assignment plainly and add: buyers should verify with the district.

Fair housing rules, required:
1. Describe the property, its features, and its amenities only. Never describe people, never describe who should live in this home, and never describe the character of the neighbours.
2. Never state or imply a preference, limitation, or discrimination based on race, color, national origin, religion, sex, familial status, disability, or any other protected class.
3. Do not use phrases that name a type of buyer or household. If you are tempted to write who a space suits, rewrite it as the feature itself.
4. Do not name a place of worship as a selling point.
5. If a fact a description needs is missing, do not guess it. Write [NEEDS AGENT INPUT: describe what is missing] in that spot instead.

Keep the tone plain and specific. Specific beats superlative: name the feature instead of calling it stunning.

The bracketed gap markers are the most valuable part of this prompt in daily use. After a listing appointment you can dictate rough notes, and you get back a clean description that shows exactly which facts still need your input, instead of a smooth draft with an invented feature buried in paragraph two and a phrase about people that should never reach the MLS.

How should you invite people to an open house, and remind them?

Invite and remind with the facts a visitor actually needs: the property, the date, the exact hours you supply, where to park or enter if you supply it, and how to ask a question. The KONAH guide notes that fair housing principles apply beyond the listing itself, to invitations, posts, captions, and how ads are delivered, so this prompt keeps the invitation about the property and the event, open to anyone, with no language about who the home is for. Write the invitation once, then run the same facts again for a short reminder the day before.

You are writing open house messages for [agent name] at [brokerage name].

Open house facts:
Property: [address and property type]
Date and hours, exactly as scheduled: [date], [start time] to [end time]
Hosting agent: [agent first name and phone]
Parking or entry facts we supply: [facts, or not supplied]
One or two property features to name, from our listing facts: [features exactly as recorded]
How to ask a question or request a different time: [reply route or phone]

Write two messages:
1. An invitation, under 90 words, for email or social posting: name the property, the date, and the exact hours, name the features supplied, and end with the question route.
2. A reminder, under 50 words, for the day before: repeat the property, date, and exact hours, add the parking or entry facts if supplied, and close with the hosting agent's name and phone.

Rules:
1. Use the exact date and hours from the facts every time, never a different window.
2. The invitation is for the property and the event. Do not describe who should attend, who should live in the home, or the character of the area's residents, and do not state any preference based on any protected class.
3. Do not invent features, prices, refreshments, incentives, or parking facts that are not supplied above.
4. Friendly and brief. No hype and no claim that the home will sell quickly.

Save both versions with your other templates. The shape of a good invitation does not change from listing to listing, only the facts do, so this is another prompt you run rarely and reuse often.

How do you ask for showing feedback without leading the answer?

Ask soon after the showing, keep it short, and ask about the property and the price reaction in neutral words. Showing agents will answer a three question message when they will ignore a long survey, and their answers are useful only if the questions did not push them toward a polite one. This prompt writes that request, plus a shorter second attempt, and it tells the model to summarise returned feedback without softening it, because a seller needs what buyers' agents actually said, in their words, about features and facts you supply.

You are writing a showing feedback request for [agent name] at [brokerage name].

Showing facts:
Property: [address and property type]
Showing date and time: [date and time as recorded]
Showing agent or brokerage, as recorded: [name or unknown]
List price as recorded, if we want a price reaction question: [price exactly as recorded, or write do not ask about price]
Any specific feature we want a reaction to, from our listing facts: [feature exactly as recorded, or none]

Write two messages:
1. A first request, under 70 words, for text or email, with three short questions: what stood out about the property, what held it back, and, only if a price was supplied above, how the price landed against what was shown. Leave space for a reply in their own words.
2. A second request, under 40 words, for the next day if there is no reply: repeat the three questions in one line each, and thank them for the showing.

Then write a summary format we can use when replies come back, with three headings: What stood out, What held it back, Price reaction. Under each heading, the rule is: quote or closely restate only what the showing agent actually wrote. Do not soften, invent, or add feedback that was not given.

Rules: do not invent showing details. Do not describe buyers or their reasons in terms of any protected class. If returned feedback mentions a protected class, flag it under a heading Flag for agent review instead of passing it to the seller as a reason.

That last rule protects the seller conversation. Feedback is about the property, its condition, its layout, and its price. Anything else gets flagged for you to handle, not repeated as if it were a market fact.

What should a market update to your sphere list actually say?

It should say what changed in a named area, using only market data you supply, and what that change could mean for a homeowner deciding on timing, without inventing a value for anyone's home. A sphere guide from MyMorningFive, citing the National Association of REALTORS 2026 Member Profile, reports that the typical REALTOR earned 28 percent of their 2025 business from past clients, and, citing the 2025 Profile of Home Buyers and Sellers, that 43 percent of buyers found their agent through a referral and 18 percent used an agent they had worked with before. That is why the update matters: your next listing is more likely to start with someone who already knows you than with a stranger. The same guide recommends several genuine personal touches through the year, supplemented by useful market communication, rather than a broadcast alone. This prompt writes the broadcast layer, from your data, with the valuation rule built in.

You are writing a market update message for [agent name] at [brokerage name], to send to a sphere list of past clients and contacts who agreed to receive it.

Market data we supply, use only this data:
Area covered: [neighbourhood, town, or market area, exactly as named in our data]
Period covered: [period exactly as recorded]
Figures we supply: [paste each figure with its source note exactly as recorded, such as number of sales, median sale price for the area and period, average days on market, or inventory count, or write none supplied]
One local fact we supply: [a recorded local fact relevant to homeowners, or none supplied]
Sender and contact route: [agent first name, brokerage, phone or reply route]

Write the update in this order:
1. A subject line or opening line that names the area and the period.
2. A short summary, under 120 words, that reports only the figures supplied above, each with the period it covers.
3. One short paragraph on what those figures could mean for a homeowner thinking about timing, written as a general observation, not advice to a named person.
4. A closing line that invites a reply with a question, and names the contact route.

Rules:
1. Never estimate or state the value of any individual home. Never invent a figure, a trend, or a comparison that is not in the data supplied above. If a figure is missing, write [NEEDS AGENT INPUT: name the missing figure] instead of filling the gap.
2. Report area figures exactly as supplied. Do not round them into a different claim and do not compare them to a period that was not supplied.
3. Describe places and property facts only. Never describe the people who live in an area, and never state a preference based on any protected class.
4. No pressure, no claim that now is the best time to sell, and no promise about future prices.

Because the prompt locks every figure to the data block at the top, whoever helps you send the update can draft it without a fresh risk each month. Still read the final version yourself. Market wording is a trust decision, not just a writing task.

How do you check in with a past client without sounding like a prospecting call?

Check in because you remember something specific, not because a calendar says to prospect. The MyMorningFive sphere guide makes that test plain: reach out because you remember something, ask about the person before talking about the market, use the channel that fits the relationship, and do not make every touch a request for business or referrals. It suggests recent clients get a check-in after closing, around their first weeks in the home, and at meaningful homeownership milestones, and that past clients get several genuine personal touches through the year. This prompt turns your notes about one person into that message, one person at a time.

You are writing a past client check-in for [agent name] at [brokerage name]. This is a personal message to one person, not a broadcast.

Facts from our notes:
Client: [client first name]
When they closed, and on what property: [closing month and year, property area or type]
What we remember from the last conversation: [note in our words, such as a renovation they planned, a new job, or a question they asked]
Milestone or reason for reaching out now, if any: [home anniversary, first weeks in the home, a note they sent us, or no specific milestone]
Channel: [call script, text, or email]
Sender: [agent first name]

Write the message in this order:
1. Open with the specific thing we remember, in plain words.
2. Ask one genuine question about how they are doing or how the home is working out.
3. Only if it fits the notes above, add one useful homeowner item: [item we supply, or write none]. Do not invent an item.
4. Close warmly, with no request for business or referrals in this message.

Rules: under 80 words for text or email, or a call opening under 40 words if the channel is a call. Do not invent facts about their home, their life, or the market. Do not mention value or price. Warm and brief, in the voice of [agent first name].

Run this prompt per person, not per list. Five thoughtful contacts built this way are the unit the sphere guide recommends as a sustainable starting point for a solo agent, and each one takes minutes once your notes are in the brackets.

When and how should you ask for a review?

Ask after a closing, when the work is done and the client can judge it, and make the ask effortless. The review request prompt below writes your request, not their review, and that line is firm for the same reason it was firm on page one of this library: genuine reviews only, never bought, never traded for anything, and never written by the model for the client to post as their own words. The tools that run this workflow are compared in AI Tools for Review Requests (2026).

You are writing a review request for [agent name] at [brokerage name].

Facts:
Client: [client first name]
What just closed: [a purchase or a sale, property area or type, closing month]
Person asking: [agent first name, the person the client worked with most]
Our direct review link: [paste the real link from our profile]
Day we closed: [date]

Write two messages:
1. A text of two or three lines: thank them by name, name the closing in a few words, say a review would mean a lot, and paste the link exactly as given.
2. One follow-up message for a few days later, under 40 words, for someone who said they would leave a review and has not yet. Warm, brief, no guilt, and no second reminder after this one.

Rules: do not write a suggested review for the client to copy. Do not offer anything in exchange for a review. Do not claim a star rating, a review count, or a ranking for our business. Keep both messages in the voice of the person asking.

That first rule, never drafting the client's words for them, protects you twice. It keeps you inside review platform rules, and it keeps your reviews sounding like your clients, which is why future buyers and sellers trust them in the first place.

How do you adapt a prompt for your market, your brokerage, and your voice?

Change one layer at a time, and test on a real past inquiry or listing before you trust a new version. Start with the market layer: swap the bracketed area names, the property types you actually handle, and the words your clients use, because a downtown condo inquiry and a rural land inquiry should not produce the same first sentence. Next, the brokerage layer: your brokerage name, your showing and offer instructions as your brokerage states them, your review link, and your callback promise, because a prompt that says you reply fast means nothing until the promise has a number on it. Last, the voice layer: read the output out loud, and anywhere it sounds like a brochure instead of you, replace that line in the prompt itself with an instruction in your own words, such as keep sentences short.

Keep the fair housing rules and the valuation rule untouched in every version you adapt. Those two rule blocks are not voice, they are guardrails, and adapting a prompt is never a reason to soften them. Keep a before and after example for each prompt you rely on: one real set of notes, and the message you actually sent after editing. When a prompt drifts, and they do drift as models update, that pair shows you in one minute what changed. Save the working versions where you can reach them, not in one lost chat thread. That storage problem is exactly what prompt tools solve, and it is why this directory built the AI Prompt Tools category in the first place.

Here is how the tools in that category fit a realtor's week, based on their listing pages in this directory. PromptBase is a marketplace for buying and selling prompts, for an agent who would rather start from a proven prompt than a blank page, with a free section to try the format first. FlowGPT is a community driven platform for sharing and discovering prompts, organized around everyday jobs including Business and Marketing categories, useful when you want to see how other operators phrase the same request. Snack Prompt is a prompt marketplace with a browsable catalog across more than 100 topic categories, including a Real Estate category named on its listing page, and a Chrome extension that puts its library inside ChatGPT, which suits an agent who lives in the chat window between showings. AIPRM is a browser extension layered on ChatGPT with a public prompt library and a private library for your own prompts, organized around jobs such as Marketing, Sales, and Customer Support, which maps closely to the lead, listing, and review work on this page. PromptFolder is a prompt manager that saves prompts in folders in the cloud, works inside ChatGPT through its Chrome extension, and lets saved prompts carry variables that clear when the prompt is loaded, which is the bracket pattern this page uses, handled by the tool instead of by hand.

You do not need all of them. Most solo agents and small teams need one place to save their own proven prompts, and one place to look when a new job appears and no template exists yet. Pick the pair that matches how you already work, load the eight prompts from this page into it, and edit them there as your wording improves.

What should you check before any AI drafted message goes out?

Check five things, every time, in this order. Facts first: names, dates, times, prices, and property facts against the source notes, because a fluent message with a wrong open house hour is worse than a plain message with the right one. Fair housing second: read every sentence and ask whether it describes the property or a type of person, and cut or rewrite any line that describes people, states a preference, or characterises who lives in an area. Value third: any value, price, or market figure in the message must trace to data you supplied, such as your recorded list price or the comparable data in the prompt, never to the model's sense of what a home should be worth. Promises fourth: anything the message commits you to, such as a callback time, a showing time, or a next step, must be a promise you will actually keep. Voice last: read it once as the client, and cut any line you would not say out loud in their kitchen.

That check takes about a minute per message, and it is the real system on this page. The prompts do the drafting. You keep the facts, the fair housing line, the numbers, and the final say. Agents who hold that line get the speed the research points to without handing their client relationships, or their licence, to a model. Agents who skip the check eventually send one confident, wrong message, and spend a month earning trust back.

Sources named on this page, checked November 13, 2026: Pinova, What Is Speed to Lead and Why It Is the Most Important Metric in Real Estate, summarizing the Oldroyd, McElheran, and Elkington study in Harvard Business Review, March 2011, and Inman's 2025 Real Estate Technology Survey, on response time and lead qualification; KONAH Media Group, Writing Listing Descriptions That Follow Fair Housing, on describing the property and not the buyer, the federally protected characteristics, and factual school assignments with a note to verify with the district; MyMorningFive, How Often Should Agents Contact Their Sphere, citing the National Association of REALTORS 2026 Member Profile and the 2025 Profile of Home Buyers and Sellers, on past client and referral business and on genuine, remembered touches over prospecting messages; and the listing pages for PromptBase, FlowGPT, Snack Prompt, AIPRM, and PromptFolder in this directory's AI Prompt Tools category, for what each prompt tool does, including the Real Estate category named on the Snack Prompt listing page. Demand evidence for this topic is a search proxy, based on the real estate lead follow up, listing description, open house, and sphere guides ranking for these jobs, not a direct Google Trends read.

Frequently Asked Questions

Can I use these prompts in ChatGPT, Claude, or Gemini without changing them?

Yes. Every prompt on this page is plain text with bracketed fields, so it runs as written in ChatGPT, Claude, Gemini, or a prompt tool extension. Fill in every bracket before you send it, because a leftover bracket is the model telling you a fact is missing. If one model writes longer answers than you want, add a word limit line to the prompt rather than editing every output by hand.

What is the fair housing rule I should keep in every listing prompt?

Describe the property and the facts, never people or protected classes. The listing description prompt on this page carries that rule in full: describe features and amenities only, never who should live in the home, never the character of the neighbours, and never a preference based on race, color, national origin, religion, sex, familial status, disability, or any other protected class. Keep that rule block unchanged when you adapt the prompt, and check the draft against it before anything is published. This is general information, not legal advice.

Can a prompt estimate what a seller's home is worth?

No, and the prompts on this page are written to refuse that job. The seller follow-up and market update prompts instruct the model to use only data you supply, such as comparable sales or area figures you paste in exactly as recorded, and to leave a bracket asking for your input when that data is missing. A value is your professional judgment, built from data you can show. It is never a gap for a model to fill.

What should I never paste into a prompt?

Never paste anything you would not want stored outside your office: full payment details, lockbox or access codes you have not been given permission to share, or private client information beyond what the message needs. Names, property facts, and dates are usually enough to draft a message. When in doubt, draft with initials and add the full details yourself in the final message.

How many of these prompts does a solo agent actually need?

Start with three: the new buyer lead reply, the listing description, and the past client check-in. Those three cover the start of a client relationship, the core marketing job, and the repeat and referral business that the sphere research on this page points to. Add the seller follow-up, open house, feedback, market update, and review prompts one at a time as each new template proves itself on real work.

How do I keep my prompt library from turning into a mess?

Save prompts in one place you can reach from any device, name each one by the job it does, and keep one before and after example with it: the notes you pasted in and the message you actually sent. Review the library once a season, delete versions nobody uses, and update the brokerage name, review link, hours, and sender names in the templates before your busy season starts.