The Real Estate AI Cheat Sheet

The problem

This book gives you 28 ready-to-paste AI prompts covering the six situations that eat an agent’s week — listing descriptions, client emails, social content, market analysis, lead follow-up, and showing prep — cutting a 45-minute listing description down to about four minutes.

You have a 9 a.m. listing presentation, two showings before lunch, an offer to counter, and three lead emails sitting unanswered since yesterday. The listing description for the new condo on Maple is still in your drafts — you started it last night and the prose came out generic. Three bedrooms, two baths, updated kitchen, great neighborhood. Every condo in the MLS says that. You know yours is different, but at 10 p.m. you could not find the words. Most agents end up with one of two bad habits: paste a template into every listing and hope the photos do the heavy lifting, or spend forty-five minutes writing each description and fall behind on the pipeline. Generic “write me a real estate listing” output reads like a bot. The prompts in this book are the ones that come back sounding like a working agent who actually walked the property.

This walks through what changes when you have a field-ready prompt library for the six situations that eat your week.

What most people get wrong

Mistake one: feeding AI three lines and expecting a finished listing. A prompt like “write me a 300-word luxury listing for a 4-bed home in Bend, Oregon” gets you generic real-estate sludge. The output mentions a “spacious living area” and “ample natural light” and the kind of phrases that compete with thousands of identical listings. The reason: the model is filling in everything you did not tell it. You have to feed it the property’s actual differentiating details — the angle of the morning light, the fact that the primary closet has a window, the school district’s specific reputation — and the buyer persona you are writing to. A starter-home buyer reads different cues than an investor. The same property needs two different descriptions, not one. Treating the prompt as fill-in-the-blanks is the move.

Mistake two: writing every email from scratch when you have already written the right one. Most agents have a “great” version of every email type sitting in one of three places: their sent folder, a stickied draft, or their head. The new-lead first response that converted last month. The post-showing follow-up that opened the conversation that led to the offer. The follow-up at day seven for a silent lead. AI’s value is not replacing your judgment — it is letting you specify the proven structure once and getting variation on it forever. The agents who scale this are not writing better prompts. They are writing the same prompts more consistently.

A working approach

The book is built around chapters that match the six situations agents actually face. Each one starts with the prompt, then the customization brackets, then the pro tip that makes the output usable.

Here is the listing-description prompt for a luxury property, exactly as it appears in chapter 2:

Write a luxury real estate listing description for:

Property: [address]
Type: [single-family home / condo / townhouse]
Beds/Baths: [number] bedrooms, [number] bathrooms
Square footage: [number] sq ft
Lot size: [number] acres/sq ft
Key features: [list 5-7 standout features]
Recent upgrades: [list any renovations]
Neighborhood highlights: [schools, parks, dining]
Price: $[amount]

Style: Sophisticated but warm. Lead with the most impressive feature.
Use sensory language (what you see, hear, feel). Mention natural light
and flow between spaces. End with a call to schedule a private showing.
250-350 words.

That alone gets you a usable draft. The pro tip is the part that makes it sound like you walked the property: add the line “Write as if describing this home to a buyer who has already seen five similar properties today and needs a reason to remember this one” to the end. That single sentence forces the AI to find differentiating details instead of generic praise. The output stops mentioning “ample natural light” and starts mentioning that the primary suite faces east, which means coffee on the deck at 7 a.m. without the afternoon glare.

The same pattern repeats across the other chapters. The new-lead response prompt limits the email to one specific question — because three questions in the first reply is the most common reason leads ghost. The seller weekly update prompt asks for one data point per paragraph — because seller anxiety scales with the gap between promised and delivered communication, and “data point per paragraph” forces specificity. The post-showing follow-up prompt requires referencing a specific detail from the showing, which lifts response rates roughly threefold over generic “thanks for coming” emails in agent A/B tests.

The book covers six situations:

1. Listing descriptions — luxury, starter home, investment, fixer-upper, social media teaser
2. Client emails — new lead, silent buyer re-engagement, post-showing, seller weekly update, offer negotiation
3. Social media content — weekly content calendar, market update video script, neighborhood spotlight, testimonial post
4. Market analysis — CMA narrative, neighborhood trend report, investment property analysis, pricing strategy memo
5. Lead generation and follow-up — nurture sequence, open house follow-up, expired listings, FSBO, past-client re-engagement
6. Open house and showing prep — talking points sheet, prep checklist, buyer consultation, virtual tour script, feedback request

Each prompt is structured the same way: paste, fill brackets, customize one or two lines for your voice, send. The first time you use one it takes ten minutes. By the fifth listing, you have a fork of it that sounds like you on a good day and the cycle time is four minutes per description.

Where this scales

The article walked through one prompt and the structure that makes the rest of them work. The book has 28 prompts across the six situations, plus a quick-reference prompt library in the bonus folder that you can keep open in a tab.

The chapters you have not seen here cover market analysis (the CMA narrative prompt is the one most agents underuse — it turns a spreadsheet into a story the seller will actually read), the lead nurture sequence (five emails over thirty days, each with a different intent), and open-house prep including a five-question feedback-request template that actually gets answered.

The point of the field manual format: you do not read it cover to cover and then start using it. You open it to the chapter that matches what is on fire today, copy the prompt, send the result, close the book. The next morning you open it to a different chapter. By month three you have your three favorites memorized and the other twenty-five are there when you need them.

Included with the book

• The complete prompt library — a single-page reference (markdown and PDF) that lists every prompt category with the bracketed fields, so you can keep one tab open while you work
• Customization pro tips — every prompt includes the one-line addition that takes the output from generic to usable

Get the full picture

Readers of this also chose

Questions readers ask

I have used ChatGPT before and the listings sounded fake. How is this different?

The difference is in the structure. Generic "write me a listing" prompts get generic output because the model has nothing specific to anchor on. The prompts in this book all share a structure: bracketed property facts, bracketed buyer persona, a style directive, and a length cap. That structure forces the output to be specific. The pro tip after each prompt is the line that takes it from passable to usable.

What if I need a refund?

Checkout runs on Lemon Squeezy. The standard refund window applies if the prompts do not work for your market. You keep the PDF either way.

Does this work with Claude and Gemini too, or only ChatGPT?

All the prompts were tested against ChatGPT, Claude, and Gemini. The output style varies slightly by model — Claude tends to be more conservative on sensory language, Gemini leans data-forward, ChatGPT splits the difference — but the prompts themselves work across all three. You do not need a paid plan to use them.

How long does it take to actually start using these?

Open the chapter that matches what is on fire today. Copy the prompt. Fill in the brackets with your property details. Paste it into ChatGPT, Claude, or Gemini. First time through it takes ten minutes. By the fifth use it takes four. The book is designed for working agents, not for reading.