中文标题待翻译

In 2024, Americans spent about US$170 billion on real estate commissions and other costs involved in transferring ownership of homes. The Bureau of Economic Analysis estimate, cited by the Federal Reserve, is equal to roughly 0.6 per cent of US GDP. This is money spent not on the homes themselves, but largely on changing who owns them.

The figure includes both new and existing homes and covers more than agent commissions. It also includes costs such as title insurance, taxes and other expenses linked to transferring ownership. AI is unlikely to remove all of this. Some parts of a property transaction exist because someone must take financial risk, establish legal ownership, inspect a building or accept legal responsibility when something goes wrong.

But a surprisingly large part of the process consists of people searching for information, comparing it, preparing documents, checking documents, answering routine questions and coordinating one organisation with another. Those are exactly the kinds of tasks AI is becoming good at. The interesting question is therefore not whether AI will replace real estate agents. It is how much of that US$170 billion exists because transferring a home still requires so much expensive human work.

The agent is really a bundle of services

A real estate agent does not perform one job. A buyer’s agent searches for properties, filters them, explains neighbourhoods, organises viewings, studies comparable sales, advises on price, helps prepare offers, supports negotiations and keeps the transaction moving. A seller’s agent values the property, prepares the listing, writes marketing material, handles inquiries, organises viewings, advises on offers and coordinates the sale through to closing.

Historically, buying all those services together made sense because performing them separately was difficult and time consuming. AI changes that economics. It can search thousands of listings, compare previous transactions, prepare marketing material, summarise contracts, organise appointments, identify unusual clauses and answer routine questions almost instantly.

That does not mean sellers suddenly stop using agents. The evidence today shows the opposite. The National Association of Realtors reported that 91 per cent of sellers still used an agent in 2025, while for-sale-by-owner transactions fell to just 5 per cent.

The first change may therefore be much less dramatic than the disappearance of agents. People may continue to use them while buying fewer services from them. A seller might use AI to check the agent’s valuation or prepare the marketing material. A buyer might use AI to find properties and compare prices, then bring in an agent mainly for negotiation and closing.

That matters because the industry can lose revenue without losing the customer. A seller who still uses an agent but pays a lower fee for a smaller package does not appear as disruption in the headline statistics. There are already early signs of this behaviour. In one 2026 survey, 23 per cent of recent sellers said they had used an AI chatbot during the sale. Among those users, 51 per cent used it to check the agent’s proposed price and 43 per cent used it to question the agent’s advice.

Also Read: Why some Gen Z graduates are skipping AI degrees entirely, and getting better jobs because of it

Search may change even faster

For years, Zillow, Realtor.com and other property portals improved home search by moving listings online. But the underlying method did not change very much. Buyers still selected filters such as price, bedrooms and location, then manually worked through the results.

AI allows something much closer to asking an experienced researcher to search on a buyer’s behalf. A buyer could ask for a three-bedroom home under a certain budget, less than 30 minutes from work, outside a flood zone, with good natural light, low renovation costs and walking access to restaurants. The AI can compare available properties against all those conditions at once.

Zillow, Redfin and Realtor.com are already moving in this direction with conversational search tools. As this improves, the value of the traditional search screen itself may decline because the difficult work of searching and comparing becomes cheap.

The valuable assets underneath it become more important. Accurate property information matters. Access to listings matters. Knowing what the buyer actually wants matters. Control over the rest of the transaction matters.

That is why AI does not necessarily make the multiple listing service, or MLS, obsolete. The MLS may remain extremely important as the trusted source of what is actually for sale. What becomes less valuable is the old interface through which people searched that information.

The same pattern runs through the rest of the transaction

Mortgage lending is another obvious target. Independent mortgage banks reported average loan production costs of US$10,936 per loan in the second quarter of 2026. Much of mortgage origination involves gathering documents, checking information, completing forms, handling exceptions and moving a regulated file through a series of steps.

Rocket reported in August 2026 that loan officers using its AI tools were handling nearly 40 per cent more clients than a year earlier. That does not mean the mortgage lender disappears. Someone still has to provide the money, assess credit risk, follow regulation and bear the financial consequences when borrowers default. AI can make the processing much cheaper without eliminating the underlying financial function.

Appraisal shows the same split. Fannie Mae already allows some eligible loans to use alternatives to traditional appraisals and estimates average appraisal cost savings of about US$550 in those cases. AI and large property databases can make routine valuation much easier, but unusual properties, physical defects and situations with limited comparable data still require real-world inspection and judgement.

Title work follows the same pattern. AI can search records, extract information from documents, prepare paperwork and identify routine problems. But someone still has to insure the title, hold money safely during the transaction, resolve ownership problems and accept responsibility if something goes wrong.

Lawyers, accountants and insurers face similar divisions. Routine document review, tax calculations, contract comparison and information gathering become cheaper. Complex judgement, legal responsibility and financial risk do not disappear nearly as quickly.

The transaction may break apart, then come back together

One effect of falling costs is that smaller specialist services become viable. A company no longer has to provide the entire real estate transaction. It might offer only listing access, only pricing advice, only negotiation support or only transaction coordination.

When the cost of running these narrow businesses falls, smaller markets become worth serving. That can produce many competitors attacking different parts of the traditional agent package rather than one new company replacing the whole industry.

Also Read: Does the iPhone show us which industries AI will destroy?

But there is an important counterforce. The transaction can also be put back together by software. Zillow is expanding beyond search into financing and closing services, while Rocket now combines mortgage origination with Redfin’s brokerage and property-search operations.

So the future may not simply be a collection of tiny services. The old human bundle can break apart while a new digital bundle takes its place. Instead of one agent coordinating everything, a software platform may coordinate the agent, mortgage, title, insurance and closing around the buyer.

Who gets the savings?

This is where the US$170 billion figure becomes important. Lower costs do not automatically mean lower prices. If many firms compete to provide a service, AI savings are more likely to be passed to consumers through lower fees. If a small number of platforms control access to customers or important property data, they may keep more of the savings themselves.

The parts of the transaction most exposed are those whose value comes mainly from expensive human work: searching, comparing, preparing documents, answering routine questions and coordinating organisations. The parts most likely to keep their value are those connected to real scarcity or responsibility: trusted data, credit risk, title insurance, physical inspection and legal authority.

Real estate agents probably sit somewhere between the two. Agents whose main value is access to listings and basic coordination are highly exposed. Agents who are trusted negotiators or who handle difficult transactions may become much more productive rather than disappear.

The same pattern applies across the rest of the industry. Routine mortgage processing loses value faster than lending capital. Standard appraisal loses value faster than difficult physical valuation. Basic legal and tax work becomes cheaper faster than legal accountability. Insurance distribution can become more efficient while the capital standing behind the policy remains necessary.

AI is unlikely to remove the entire US$170 billion Americans spend transferring homes. It does not need to. If even a modest share of that cost exists because people currently perform work that software can do much more cheaply, billions of dollars are available to move between households, agents, brokers, lenders, lawyers, title companies and software platforms.

Read the full report, “Can AI Cut the US$170 Billion Cost of Buying and Selling Homes?” here, and the book Abundance Economics: The Economic Mechanisms of the AI Age here.

For more information on this proprietary research framework analysing AI-driven economic change, please visit here. 

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post Can AI cut the US$170 billion cost of buying and selling homes? appeared first on e27.