AI for real estate agents: faster follow-up, careful wording

Real estate is a relationship business run out of an inbox, a phone and a CRM. That is a good setup for AI, because the repetitive parts, including follow-up, listing text and paperwork summaries, are separable from the part clients pay you for: knowing the market and guiding a decision.

Where AI helps agents

Follow-up that actually happens. Most leads go cold because follow-up is inconsistent. Paste in your showing notes or the inquiry, add two lines about the client, and get a draft reply in your own voice. The Gmail guide shows how to set up a workflow that triages your inbox, drafts replies, and flags action items.

Listing descriptions. Give the assistant the verified facts: square footage, updates, lot, neighborhood features you can document. Ask for three versions with different tones. It writes well from facts. It writes badly, and sometimes falsely, when it fills gaps, so instruct it to use only what you provide.

Paperwork summaries. Purchase agreements, inspection reports and HOA documents are long. Ask for a list of dates, contingencies, deadlines and open items with the location of each in the document. See can AI review contracts. Summaries help you prepare; they do not replace your broker, attorney or the actual document.

Cleaner records. Convert messy call notes into a consistent CRM entry with the next step and date.

Where AI does not help

• Local market truth. It may state prices, trends or school information that are outdated or invented. Use MLS and official sources, and give the assistant the numbers you want it to describe.
• Valuation. A price opinion belongs to you and your comparable data.
• Legal and disclosure questions. Requirements vary by state. Route them to your broker or an attorney.

Risks specific to real estate

Fair housing. Listing and marketing text must not express preference or limitation based on protected classes. An AI trained on internet text can drift toward phrases like “perfect for young families” or “walking distance to church” that steer buyers. Describe the property, not the ideal occupant. Always review the output, and ask the assistant to flag risky phrases as a second check, not a substitute for your own read. Follow your brokerage’s compliance guidance and the fair housing rules that apply where you work.

Client privacy. Financial details, offer terms and personal circumstances are sensitive. Read whether it is safe to put company data in ChatGPT and keep names and identifying details out when the task does not need them.

Accuracy in advertising. Every claim in a listing, from “recently renovated” to square footage, needs support. A model will not know what you did not tell it.

Starter prompts

Listing draft:

Write three listing descriptions of about 120 words each, in
warm, friendly and concise tones. Use ONLY the facts below. Do not
add features, neighborhood claims or anything about who would
like the home. Describe the property, not the buyer.

Facts: [paste your verified facts]

Follow-up:

Draft a short reply to this buyer inquiry. I am [name]. Answer
their question using only the details I provide, suggest two
times for a showing, and keep it under 90 words.

Inquiry: [paste]
My details: [paste]

Your first week

• Day 1: Take five recent follow-ups you wrote and try the follow-up prompt. Note where the draft needed edits.
• Day 2: Draft one listing from verified facts. Check every sentence against the property and against your brokerage’s fair housing guidance.
• Day 3: Summarize one past contract into a dates-and-contingencies checklist.
• Day 4: Set up a consistent format for CRM notes and convert three days of notes.
• Day 5: Decide what data never goes into an AI tool, and write it down.

A worked example: from showing to follow-up in five minutes

After a showing, you jot four lines: the buyer wants a bigger yard, dislikes the kitchen, is pre-approved, and can move by spring. Paste those into the assistant along with the property facts and ask for a follow-up email under 100 words that references the yard and the timeline, offers two other listings you name, and asks for a time to talk. Then ask for a CRM entry with stage, budget signal, timeline and next step.

You read the email, correct anything you would not say, and send. The CRM entry takes ten seconds to paste. Over a month, that is the difference between every buyer getting a same-day follow-up and only the ones you had time for.

Common mistakes to avoid

• Letting it describe the neighborhood. It will make claims about schools, safety or demographics that you cannot support and that can raise fair housing concerns.
• Publishing first drafts. Listing text written from a thin fact sheet sounds like every other listing.
• Pasting offers and financials with names attached. Strip what the task does not need.
• Assuming a contract summary replaces your broker’s review.

What the sources say

Labor data for agents and brokers, and HUD fair housing guidance that covers AI-assisted advertising.

• The Bureau of Labor Statistics projects overall employment of real estate brokers and sales agents to grow 2 percent from 2025 to 2035, slower than the average for all occupations.
• The Bureau of Labor Statistics says every state requires real estate brokers and agents to be licensed.
• The Department of Housing and Urban Development guidance says intentional discrimination can include using a protected characteristic as the basis for a housing decision.
• The Department of Housing and Urban Development guidance says a housing decision made in whole or in part by an automated system, including one using machine learning or another form of AI, is treated the same way.
• The Department of Housing and Urban Development guidance says advertisers and ad platforms may violate the Fair Housing Act by selecting audiences for housing-related ads based in part on protected characteristics or proxies.

If you want a guided start

Learn AI Fast is a two-day in-person class in Boston where non-developers practice email, document and workflow automation on provided case files.

Sources

• U.S. Bureau of Labor Statistics, Occupational Outlook Handbook: Real Estate Brokers and Sales Agents: “projected to grow 2 percent from 2025 to 2035, slower than the average for all occupations”
• U.S. Bureau of Labor Statistics, Occupational Outlook Handbook: Real Estate Brokers and Sales Agents: “Every state requires real estate brokers and agents to be licensed”
• U.S. Department of Housing and Urban Development, Guidance on Application of the Fair Housing Act to the Advertising of Housing through Digital Platforms (April 2024): “Intentional discrimination can include using a protected characteristic”
• U.S. Department of Housing and Urban Development, Guidance on Application of the Fair Housing Act to the Advertising of Housing through Digital Platforms (April 2024): “This is true even if that decision is made in whole or in part by an automated system, including a system using machine learning or another form of AI”
• U.S. Department of Housing and Urban Development, Guidance on Application of the Fair Housing Act to the Advertising of Housing through Digital Platforms (April 2024): “Advertisers and ad platforms may violate the Act by segmenting and selecting audiences for housing-related ads based in part on protected characteristics or proxies”