AI for Real Estate Marketing: What Agents Hand It and What They Keep

Nearly half of US agents now use AI every day or every week, and most of what they hand it is writing: listing descriptions first, then social posts and follow-up emails. What they keep is the responsibility for what goes out, so every draft still passes through the agent before a buyer reads it.

From Trying AI to Using It Every Week

In NAR's 2025 technology survey, taken in July 2025, 32% of agents had not actively tried AI for their business, and only one in five used it daily. By NAR's 2026 REALTORS® Technology Report, a few more agents used it daily, another quarter used it weekly, and only 12% said they weren't using it and had no plans to. The two surveys asked slightly different questions, but they point the same way.

AI has gone from something agents talked about to part of an ordinary working week, so what is left to decide is which jobs to give it and how closely to check each one.

Words Before Pictures

Writing listing descriptions was the top use in 2026, named by 75% of the agents who use AI. More than half used it for social media posts and about half for emails and follow-up. Smaller groups, around three in ten, used it for market summaries or to draft marketing in their own tone, and a little fewer to review or summarise documents. NAR's guide to AI in real estate puts generative AI the same way, as a tool for listing descriptions and marketing content.

Every one of those jobs produces text, and for a single listing much of it repeats the same facts. The description, the Instagram caption and the email to a buyer who asked about the house all carry the price, the bedrooms and the area. Each draft is only as accurate as the details it started from, and a wrong figure turns up wherever the same details were used.

Pictures come further down. NAR names photo enhancement among the AI tools reshaping real estate, but none of the six uses in its 2026 list is an image task. AI photo editing covers everything from correcting the light in a room to adding furniture or putting a blue sky over a grey one, and the rules treat those two ends differently.

What Still Goes Out Under Your Name

A draft becomes the agent's own statement once it is published. Article 12 of NAR's 2026 Code of Ethics asks REALTORS® to be honest and truthful in their real estate communications and to present a true picture in their advertising and marketing. NAR's account of AI's risks puts it in practical terms: agents have to oversee their AI tools and respond when one gives inaccurate, misleading or discriminatory results.

Checking is therefore part of the job. Read every draft against the listing: the price and its currency, the area and its units, the room count, and any claim about the neighbourhood, which a generated draft can make without anyone having verified it. Fair housing rules apply to that text as they do to anything else in an ad, since 24 CFR 100.75 bars words, phrases or photographs that indicate a preference for or against buyers by race, colour, religion, sex, disability, family status or national origin.

For photos, California's law on digitally altered images has applied since 1 January 2026: an ad with a photo that has had something added, removed or changed must say so beside the image and link to the original, while changes to lighting, white balance or exposure don't count. Keeping the original of every edited photo, and labelling the edited one, gives you what both parts of that rule need.

Time Back, and What It Costs

In the 2026 report, 81% of agents said saving time was their main reason for taking up new technology, up from about two thirds a year earlier, and improving the client experience came second. NAR's Deputy Chief Economist, Jessica Lautz, put the payoff this way: "When routine work gets handled faster, agents have more time for the guidance, negotiation, and steady advice that clients count on through a major financial decision."

The obstacles to new technology were the learning curve, named most often, and cost, named by 59%. A little over half of agents said AI had a positive effect on their business. Checking drafts takes back some of the time saved, and it goes quicker when every draft starts from details that were right in the first place.

One Project's Facts Behind Every AI Draft

In Giraffe360, the drafts start from the project the listing was captured into. Its Property Data holds the address, price, bedrooms, bathrooms, area and the contact shown on the material, along with a brand accent colour, all entered once when the project's marketing is set up and corrected there later.

Marketing content in Content Studio turns those details and the listing photos into a flyer, a social media image, an open house graphic or a newsletter graphic. Before anything is generated, its Data panel shows the content it will use and the images, with the Hero photo marked, and any field can be corrected there. The finished design opens in Preview, where you read the text and check the price and area units before Save to content.

The Property Description drafts the listing text. You choose the Writing style and Description type, switch off rooms you don't want mentioned on the Information tab, review the nearby places on the Location tab, then select Generate description. Claims in the draft about the location may not have been verified, so check those before the text goes into a listing.

For photos, the still photo editor can add a disclosure watermark, with Digitally Altered as its default label and a position you choose, and Apply to copies the setting to other photos. AI enhanced results are saved separately from the originals, so the unaltered photo is there when a rule asks for it.

Spending on AI is set by the organisation's admins and managers in Billing, under Manage AI credits. Paid AI enhance credit access names the colleagues allowed to use paid AI credits, the Spending limit caps paid use across the organisation each calendar month, and Usage breakdown shows who used credits and on which projects.