An AI real estate agent is software that guides a home buyer through the search, the viewings, the offer and the purchase, doing much of the work that a human buyer's agent does. It is also called an AI buyer's agent. It is offered mainly in America, where a buyer's agent has traditionally been paid a commission of about 2% to 3% of the price of the home.
In America, a buyer usually has their own agent, separate from the seller's agent. The buyer's agent finds homes, books viewings, advises on the price to offer, negotiates with the seller's agent and follows the paperwork until closing. An AI agent aims to do most of this work in software. The buyer then pays less than the usual commission, or receives part of it back.
How does it work? The AI agent is built on a foundation model, a large AI model trained on huge amounts of text, which can hold a conversation. The buyer chats with it in an app. It asks about her budget, the areas she likes, schools and her commute. It searches the listings on property websites and explains each home. It books viewings and prepares the offer. A licensed human agent still checks and signs the documents, because American states require a licensed agent for these steps. The AI agent is part of the agency's brokerage software, which tracks the deal to closing.
Take, for example, a first-time buyer in Phoenix, Arizona, with a budget of $350,000. In the first chat, the AI agent finds 38 homes that match. It narrows these to 9 by her commute and the school ratings she cares about. It books 4 viewings in one weekend. She chooses a home listed at $355,000. The AI agent shows her that 5 similar homes nearby sold for $340,000 to $352,000, and suggests an offer of $345,000. The seller accepts $348,000. A licensed agent reviews the contract and sends it for electronic signature. A 2.5% commission on $348,000 would be $8,700. Instead, the agency returns part of that commission to her as a rebate.
An AI that searches, compares and acts on a buyer's behalf is an example of agentic commerce. The difference here is the size of the purchase. A home is the largest thing most people ever buy, so the AI prepares the decisions, and the buyer and a licensed agent make them.
What can go wrong? An AI agent cannot see what a person notices on a visit, such as damp walls or a noisy road. So it does not replace the viewing or the professional inspection of the home.