How to Automate Your Real Estate Business With Claude: The Full System

Most agents who try AI stop after one good listing description. The prompt worked, they saved it in Notes somewhere, and it never became a habit. Three weeks later they're back to writing follow-ups from scratch at 9pm, because nobody ever turned "a prompt that worked once" into a system that runs every week without being rebuilt from memory each time.

This guide is that system, built specifically for Claude, not for a generic AI chat window. It covers the ten pieces of writing a solo agent repeats every week, a business context file that makes every output sound like your brokerage instead of a generic template, a fixed weekly cadence for when each piece runs, a compliance check most AI guides skip entirely, and what to do when Claude gets it wrong. By the end you'll have a working system, not just a list of prompts to lose in a chat history.

Everything here works in Claude.ai (a Project) or Claude Code (a CLAUDE.md file), and the underlying prompts are also in the Real Estate prompt pack as a standalone reference if you just want to copy one at a time.

What you're building

Two pieces, working together.

The first is a business context file: a short document that tells Claude who you are, so it stops writing like a generic AI and starts writing like your brokerage. Without this, every output sounds the same regardless of which agent is using it. With it, Claude knows your market, your tone, and the details worth repeating.

The second is a weekly cadence: the same ten prompts from the prompt pack, placed on fixed days instead of used whenever you happen to remember them. A prompt you use once is a trick. A prompt on a fixed day is a system.

Open houses and objection handling aren't on the calendar. They run the same day the open house or the stall happens, while the details are still fresh enough to write about specifically.

What you need

• A Claude account. Either works: Claude.ai with a Project (simpler to start), or Claude Code with a CLAUDE.md file in a folder (more setup, but it remembers everything across sessions without you re-pasting context, and can eventually read your actual CRM exports and MLS data directly).
• Ten to fifteen minutes to write your business context file once. This is the only setup step. Everything after this is copy, paste, review, send.
• Nothing else. No CRM integration, no API keys, no code. If you later want Claude to read your CRM exports directly, Claude Code can do that too, but it's not required to get value from this system.

Step 1: Write your business context file

This is the step that separates a system from a party trick. Paste this into a Claude Project's custom instructions, or save it as real-estate-context.md if you're working in Claude Code.

I'm a real estate agent. Here's my context:

Brokerage: [your brokerage name]
Market: [city/region, and one sentence on what's distinctive about it
  right now, e.g. "inventory is tight, most listings get 3+ offers"]
Specialty: [e.g. first-time buyers, luxury, relocation, investment
  properties]
Tone: [pick one and be specific: warm and conversational / direct and
  data-led / formal and detail-oriented]
Things I never say: [e.g. "won't last long", "must see", any phrase
  that makes you sound like every other listing]
Compliance: All listing and marketing copy must comply with the Fair
  Housing Act. Never reference or imply preference for any protected
  class (race, color, religion, sex, familial status, national origin,
  or disability). Describe the property and the neighborhood's
  amenities, never the type of person who "belongs" there.

Fill in the brackets with real details, not placeholders. "Warm and conversational" produces different output than "direct and data-led," and Claude will actually follow whichever one you pick, so it's worth being specific rather than generic. The compliance line isn't optional. More on why in a few sections.

If you're using Claude Code, put this in a CLAUDE.md at the root of a folder you work from. Claude Code reads that file automatically at the start of every session in that folder, so you never re-paste it. If you're using Claude.ai, paste it once into a Project's custom instructions and every conversation inside that Project inherits it.

Step 2: Load the ten prompts

The prompt pack has all ten, each with a copy button. In Claude Code, the cleanest setup is a prompts.md file in the same folder as your CLAUDE.md, so both are always available together. In Claude.ai, add them as Project knowledge, or just keep the pack open in a tab and paste from there.

Either way, you don't need to memorize them. You need to know which one runs on which day, which is the next section.

Setting this up in Claude.ai versus Claude Code

Both work. The difference is where the context lives and how much it does on its own.

Claude.ai (a Project). Create a Project, name it something like "Real Estate Assistant." Paste your business context file from Step 1 into the Project's custom instructions. Add the ten prompts as Project knowledge files, or just keep the prompt pack page open in a browser tab. Every new conversation started inside that Project automatically has your context loaded, no re-pasting. This is the right starting point if you've never used Claude for work before, since there's nothing to install and no files to manage.

Claude Code. Create a folder on your computer, something like real-estate-ai/. Save your business context file as CLAUDE.md in that folder, Claude Code reads this automatically at the start of every session there. Save the ten prompts as prompts.md in the same folder. The real advantage over Claude.ai shows up once you want Claude working with actual files: drop your weekly CMA exports, MLS data pulls, or a running log of past listings into that same folder, and Claude Code can read and reference them directly instead of you copying numbers in by hand each time. It's more setup for less daily typing, which is worth it once this system is a genuine weekly habit rather than something you're still testing.

Start with Claude.ai if you're not sure yet. Move to Claude Code once you're running this every week and the copy-paste is the part that's starting to feel repetitive.

Step 3: Run the weekly cadence

Here's each day with a real worked example, not just the prompt in the abstract.

Monday: listings and price reductions

Anything that went live or got repriced over the weekend gets written up first, before email, before anything else. Writing this Monday morning, not "when there's time," is what actually gets it done, because "when there's time" reliably means Thursday.

Worked example. Your raw walkthrough notes:

612 Maple Ridge, 3bd/2ba, 1,850 sqft, built 2004. Kitchen redone 2023,
quartz counters, gas range. Big backyard, mature oak tree, deck needs
staining. 8 min to elementary school. Sellers relocating for work,
motivated but not desperate. Asking 475k.

The prompt (from the pack):

I'm listing a property. Here are my raw notes from the walkthrough and the spec sheet: . Write an MLS listing description under 200 words. Lead with the one detail a buyer would actually remember, not the generic opener ("Welcome home to..."). Use specific details from my notes, not filler adjectives. No exclamation points. End with one line about who this home is actually right for.

A reasonable Claude output, with your business context file loaded:

The kitchen was redone in 2023 with quartz counters and a gas range, the kind of update most buyers plan to do themselves within the first year. This 3-bedroom, 2-bath home at 612 Maple Ridge sits on 1,850 square feet, built in 2004, with a backyard shaded by a mature oak tree and an 8-minute walk to the elementary school. The deck needs staining, worth knowing before you budget for it. A solid fit for a family that wants the big-ticket renovation already done and is willing to handle a weekend project instead.

Notice what it didn't do: no "won't last long," no invented urgency, no claim about who "belongs" in the neighborhood. It used your actual notes, including the deck needing work, because honest listings build more trust than ones that hide the one flaw a buyer will find on the walkthrough anyway.

Tuesday: follow-ups

Everyone shown a property in the last seven days who hasn't responded gets a follow-up. This is the highest-leverage prompt in the entire pack, because it's the task agents skip most. The showing that goes quiet is where deals actually die, not in negotiation.

The prompt needs two real inputs to work well: the buyer's name and address, and whatever you remember them saying, even one sentence. "They mentioned the kitchen felt small" turns a generic follow-up into a specific one. If you genuinely remember nothing, the prompt still produces a reasonable plain follow-up, but the specific version converts better, so it's worth the ten seconds to jot down one line right after every showing while it's fresh.

Wednesday: the market update

One email, sent the same day every week, to your full list. The content changes weekly, the day doesn't. Consistency is what turns this from an email people scan into an email people expect. A market update sent whenever you get around to it reads as noise; the same update sent every Wednesday at 9am reads as a resource.

Pull whatever numbers actually moved this week (median price, days on market, inventory, one local story if there is one) and let the prompt do the writing. Lead with the single most useful number, not a summary of everything.

Worked example. This week's raw numbers:

Median price: 482k (up from 468k last month)
Days on market: 19 (down from 27)
Active listings: 34 (down from 41)
Notable: two multi-offer situations this week on sub-500k listings

A reasonable Claude output:

Homes in our area sold in a median of 19 days this month, down from 27 the month before. That's the number worth paying attention to: inventory is tightening (34 active listings now, versus 41 a month ago), and we saw two multi-offer situations this week on properties under 500k. If you've been waiting for a slower market to buy, this isn't it. If you've been on the fence about listing, the current pace of sales works in your favor. Reply to this email if you want a read on what your specific home would list for in this market.

Notice it didn't just list four numbers in a row. It picked the one that mattered (days on market dropping) and led with it, then used the rest to support two different, specific pieces of advice, one for buyers and one for sellers, instead of one vague paragraph trying to speak to everyone at once.

Thursday: seller conversations

Anything requiring a harder conversation, a price reduction, a CMA presentation, a counter-offer explanation, gets prepared with the relevant prompt before the call, not improvised during it. This is talking points, not a script to read verbatim. Claude is good at ordering the argument (data first, your recommendation second) and bad at knowing your specific seller's temperament, so read the output and adjust the tone before the call, don't paste it into a text message unedited.

Worked example. A seller conversation you're dreading:

Seller: the Whitfields, listed at 550k, 41 days on market, no offers.
Feedback from 6 showings: "overpriced" x3, "needs updating" x2, "loved
it but scared off by price" x1. Comparable sales: 3 homes sold in the
last 60 days between 505k-525k, similar size, one with a renovated
kitchen (sold at 525k), two without (sold at 508k and 512k). The
Whitfields' kitchen is original, 1998.

A reasonable Claude output:

Lead with the showing feedback, not the comps. They've had six showings and no offers, and three of six said "overpriced" without prompting, that's the sellers' own market telling them something, not just your opinion. Then bring the comps: three similar homes sold in the last 60 days, all under 525k, and the one that hit 525k had a renovated kitchen theirs doesn't have. Their kitchen is original from 1998, which puts them closer to the 508k-512k range on comps alone, before accounting for 41 days of sitting on the market working against them further. Recommend landing at 519k: enough of a cut to generate new showings, not so much that it signals desperation. If they push back with "let's give it more time," the honest answer is that 41 days without an offer already is more time, and the risk of sitting longer is a further price cut that looks worse than one clean adjustment now.

That's the shape every price-reduction conversation should take: their own feedback first, because a seller trusts what buyers actually said more than what their agent says, then the data, then one specific number with a reason attached, then a prepared answer for the objection you already know is coming.

Friday: social and referrals

New listings from the week get the social post prompt. Anyone who closed in the last two weeks gets the referral request. Both get skipped constantly because neither feels urgent on any given day, and both compound if they aren't done. A referral ask sent one week after closing, while the relief and gratitude of a finished transaction are still fresh, converts at a completely different rate than one sent three months later when the memory has faded into just another closed file.

As they happen: open houses and objections

The open house batch follow-up and the "we want to think about it" objection response aren't scheduled because they can't be, they run the same day the event happens. Paste your sign-in sheet notes right after the open house closes, while you can still remember which name asked about the school district and which one asked about the roof.

Worked example. A buyer just told you they want to "think about it" after seeing 612 Maple Ridge:

Buyer: the Ortegas, saw the property yesterday, liked the kitchen and
backyard, said the price felt slightly high compared to a similar home
they saw last week (unclear if that comp is really comparable). They
have a lease ending in 6 weeks, no other showings scheduled yet.

A reasonable Claude output:

Acknowledge the hesitation directly instead of pushing past it: "Totally understand wanting to think it over, it's a big decision." Then surface the one real, time-bound fact: their lease ends in six weeks, and with no other showings scheduled, the math on timeline is worth naming gently, not as pressure but as a real constraint they raised themselves. On price, ask rather than argue: "What was the comp you were thinking of? Happy to walk through how it compares, sometimes square footage or condition explains the gap." Close with one specific, low-pressure next step: a second look at a quieter time, or the seller's disclosure packet to review before deciding either way.

The pattern here matters more than the specific words: name the hesitation, surface one real fact instead of manufacturing urgency that isn't there, ask rather than argue on the price objection, and end with a concrete next step instead of a vague "let me know if you have questions."

The compliance check most guides skip

The Fair Housing Act prohibits discriminatory language in real estate marketing, and this is genuinely not optional context, it's federal law that applies regardless of intent. An AI model writing a listing description has no idea it's about to write something that implies a preference for a protected class. It will happily produce "perfect for a young professional" or "ideal for a growing family" because those phrases sound like normal marketing copy, not because it's trying to discriminate. The business context file in Step 1 includes an explicit compliance instruction for exactly this reason: it's the cheapest possible insurance against a real problem.

Even with that instruction loaded, read every listing description before it goes out and watch for two patterns specifically: any phrase describing who the buyer should be rather than what the property offers, and any phrase describing the neighborhood in a way that implies who lives there rather than what's nearby. "Close to the elementary school" describes the property. "Great for families" describes the buyer, and that's the line.

A few concrete swaps worth having memorized, since these are the phrases that slip through most often:

• Instead of "perfect for a young couple," write "two bedrooms, ideal for a home office or guest room."
• Instead of "great for families," write "walking distance to the elementary school and the park."
• Instead of "quiet, safe neighborhood," write "low-traffic street, cul-de-sac lot."
• Instead of describing what a unit lacks as a selling point (implying who it's not for), just state the physical layout: "second-floor unit, no elevator."
• Instead of "walking distance to [a specific church or temple]," write "walking distance to downtown," or name the distance in minutes rather than naming the religious institution.

The pattern across all five: describe the physical property and its measurable proximity to things, never the type of person who fits it. This isn't legal advice, and if you have real questions about a specific listing, a real estate attorney is the right call, not this article. But this is a well-known, well-established rule, and it's worth building the habit of checking for it every single time, not just when something feels obviously off.

What Claude gets wrong, and how to fix it

It invents specificity when your notes are thin. If you give it three words about a property, it will sometimes fill the gaps with generic real-estate-sounding language instead of asking you for more. Fix: feed it more detail, not less. Five real details produce a better description than fifteen generic adjectives.

It defaults to urgency language unless told not to. "Won't last long," "must see," "won't want to miss this," these are the AI equivalent of filler words, and most models reach for them by default because they're common in training data. The "things I never say" line in your business context file is what stops this. If it still slips through occasionally, just tell it directly in that conversation: "drop the urgency language."

It can't tell an honest CMA argument from a hopeful one. When you ask it to prepare price-reduction talking points, it will build the strongest case from whatever data you give it, but it has no independent judgment about whether that data actually supports the number you're hoping for. Read the comparable sales yourself before the conversation. Claude organizes the argument; you're still the one deciding if the argument is sound.

It doesn't know your market's unwritten rules. Every market has conventions Claude has no way of knowing: which neighborhoods are sensitive to certain phrasing, which school district comparisons come up constantly, what "walkable" actually means in your specific city. This is exactly what the business context file is for, and it's worth updating it the first few times you notice Claude missing something a local agent would automatically know.

Extending this to your own business

Every prompt in this system follows the same shape: give Claude real specifics, name what you don't want, ask for a length limit and a specific ending. If you want to add an eleventh prompt for something this pack doesn't cover, an open house invite, a past-client check-in, a new-agent introduction, write it the same way. The business context file does most of the real work; the individual prompts just point it at a specific task.

If you're not in real estate, the shape of this system, ten pieces of recurring writing plus a business context file plus a fixed weekly cadence, is the same one used for the recruiting, accounting, and marketing agency prompt packs as they publish. The specific prompts change. The system underneath doesn't.

FAQ

Do I need Claude Code, or does Claude.ai work? Claude.ai with a Project works fine to start. Claude Code is worth it once you want Claude reading your actual CRM exports or MLS data directly instead of you copying details in by hand every time.

Will this sound like every other agent using the same prompts? Only if you skip the business context file. The prompts are the same for everyone who copies them; your market, your tone, and your specific listing details are not. Two agents running this exact system on two different listings will produce two different outputs, because the inputs are different.

How long does this take once it's set up? The business context file is a one-time, fifteen-minute setup. After that, each individual prompt takes under a minute to run, the time is in reviewing and editing the output, not generating it, which is the point: it should save you the blank-page time, not the judgment time.

What if Claude writes something inaccurate about a property? Treat every output as a first draft prepared in thirty seconds instead of ten minutes, not a finished decision. You still need to verify every factual claim (square footage, school assignment, HOA fees) against your actual listing data before it goes out. Claude doesn't know your MLS data unless you give it to that conversation.

Does this replace my CRM's automated drip emails? No, and it isn't trying to. A CRM's automated sequence is the same three emails to everyone regardless of what actually happened at the showing. This system produces one specific, individually reasoned message per person, which is a different kind of tool for a different part of the job. Many agents run both: the CRM handles routine touch points on autopilot, this handles the writing that genuinely needs a human read of what happened.

What happens the first week I skip a day? Nothing breaks, and that's the point of building this as a cadence rather than a fragile all-or-nothing routine. If Tuesday's follow-ups don't happen, they're still worth doing Wednesday, a slightly late follow-up beats no follow-up. The value isn't in perfect adherence, it's in the writing getting done on most weeks instead of the rare week you happen to have spare time.

Can I use this for a team, not just a solo agent? Yes, with one change: give each agent their own business context file rather than sharing one. The tone, specialty, and market details should reflect the individual agent writing the message, not a single team-wide voice, otherwise every agent's client starts receiving emails that sound identical to every other agent's client, which undercuts the personal relationship the whole system is meant to strengthen.

What this doesn't replace

None of this writes an offer, negotiates a contract, or makes a pricing call. Every output here needs a human read before it goes out, especially the seller conversations, where the actual judgment call, how hard to push, when to fold, is not something a prompt should make for you.

Getting started today

Write your business context file first. It's the one step that makes every prompt after it better, and it takes fifteen minutes. Then pick one day, Monday is the easiest since it only requires a new or repriced listing to exist, and run the listing description prompt once on a real property. If the output needs one edit, you're ready for the rest of the week. If it needs five, your notes going in were too thin, not the prompt failing, so add more specifics and try again.