AI could push real estate commissions lower, Alloy Advisors says

A new research report from Alloy Advisors finds that a typical $400,000 home sale in the U.S. generates about $39,660 in transaction costs, with real estate commissions making up the majority of the friction. According to the report, AI is poised to accelerate downward pressure on the traditional 5%-plus commission model.

In “The Home Sale Transaction, Reconsidered,” authors Amit Kulkarni and Russ Cofano of Alloy Advisors, dissect the economics of a standard U.S. resale transaction in 2025 and argue that the current real estate commission structure is misaligned with the actual value delivered.

Using a $400,000 sale as a benchmark, the authors estimate a combined $39,660 in “hard” transaction costs — about 9.92% of the sale price — flowing to third parties, based on national averages and cross-market data. Sellers bear about three-quarters of that total, or $30,200, while buyers pay roughly $9,460 at closing beyond their down payment.

According to Kulkarni and Cofano, for housing professionals, the findings of the report present both a margin risk and a competitive opening: AI-enabled consumers will be able to see, question and negotiate specific line items in ways that were not possible even a few years ago.

“The real estate industry thinks that AI is just here to serve the industry and real estate professionals, but that is not the case,” Kulkarni told HousingWire. “AI is here to serve whoever wants to use it and that could mean a variety of different things for different industries, but when it comes to the real estate transaction, I think it will mean that consumers will soon find it overpriced.” 

For Kulkarni, AI is putting a point on the cost of transacting real estate as it is becoming clearer the value the human being can provide and the value AI provides. 

Commissions dominate the friction

According to the analysis, real estate commissions account for $23,000 of the $39,660 in total costs on the sample transaction — 5.75% of the sale price and 76% of all seller-paid friction. The remaining $16,660 is taken up by things like transfer taxes, owner’s title insurance, settlement fees, loan origination fees, underwriting fees, appraisals and home inspections. Additionally, the analysis argues that embedded in an agent’s commission is a “platform tax,” which it attributes to portal referral programs and MLS fees that rarely appear as standalone charges to consumers.

Post-settlement commissions have not fallen

The report directly addresses expectations that the National Association of Realtors’ (NAR) commission lawsuit settlement would compress real estate commission rates. Citing data from Clever Real Estate and Redfin, the authors note that the national average commission rose to 5.44% in mid-2025, up from 5.32% the prior year. Additionally a HousingWire survey from April 2025, found that 58.8% of agents said their buy-side commissions had not changed since the settlement went into effect in August 2024, while 11.76% reported that their commissions had increased. 

According to the report, there are several structural reasons why commission rates have not changed, including that the business practice changes did not remove seller-paid buyer commissions in practice, that 13 states and Washington, D.C., effectively ban à la carte brokerage services and the system is structured in a way that only compensates agents when a deal closes. 

Kulkarni and Cofano write in the report that these factors help explain why overall commission levels have been sticky despite significant legal and regulatory change — and why simple disclosure shifts may not be enough to move the national average.

What the agent’s work is actually worth

A central thesis of the paper is that most of the tasks historically bundled into a 3% listing commission have been commoditized by software and AI, while the remaining “human core” of the job does not scale with home price. The report separates agent tasks into two tiers:

• Tier 1 – AI-compressed tasks: Comparative market analyses, MLS entry, listing descriptions, offer modeling, transaction coordination and basic disclosure checks. Pre-AI, the report pegs their combined market value at roughly $1,500 to $3,500 per listing. With modern tools, the authors estimate the marginal cost of these services has fallen close to zero for a competent AI user, aside from $10 to $30 per deal for workflow software.
• Tier 2 – Human-value tasks: Skilled negotiation execution, emotional coaching, on-site judgment, hyperlocal knowledge and licensed fiduciary accountability. Using comparable professional service benchmarks, Alloy Advisors estimates this “human core” is worth roughly $2,000 to $6,500 per transaction, regardless of home price.

By comparison, a 3% listing commission on a $400,000 property is $12,000, and on a $1.5 million property it is $45,000, even though the underlying Tier 2 work does not increase proportionally.

Where agents continue to bring value 

According to Cofano and Kulkarni, the commission model’s indifference to skill is the heart of the problem, as consumers cannot reliably distinguish a top-decile agent from a median one before signing a contract with one, yet both typically charge the same percentage rate.

For agents, teams and brokerages, the implication is that sustained premium pricing will increasingly require demonstrable performance on the specific tasks where human skill still moves outcomes — particularly negotiation and local market insight — as AI not only automates more tasks for agents, but provides consumers with more information. 

“The more AI is used by consumers and the more information about a transaction it can provide them, becoming a real legitimate tool for them to process information as opposed to the human hand they are holding through the transaction. For the industry to not expect that to impact the fundamental economics of this, is just fantasy,” Cofano said. 

The report cites Realtor.com research from 2025, that shows that about 82% of active or potential buyers and sellers reported using AI for housing insights and respondents rated agents and AI nearly evenly on which source made them “smarter” about the market, with agents still perceived as more accurate overall. However, a YouGov survey from December 2025 cited in the paper found that 65% of Americans trust AI to compare prices on major purchases, including homes, but only 14% expressed trust in AI to act on their behalf in such decisions.

This level of information trust, according to Kulkarni and Cofano, is sufficient to put downward pressure on real estate commissions and other fees because AI can evaluate proposed terms, line items and alternatives in real time. 

“Whether they like it or not, these things are happening and they will have an impact on the structure of the business, the compensation of the business, how services are delivered and what services need to be delivered,” Kulkarni said. “Folks need to understand that this is not going to be the status quo. Things are going to change now that consumers are empowered with these tools.” 

Change is afoot 

For housing professionals, the authors say the takeaway is that near-term competitive pressure will come from AI-assisted consumers and new pricing options, long before large-scale regulatory overhaul or pure AI listing platforms reshape the landscape.

“People have always said they wanted a better way to transact real estate, but there really has not been a viable alternative that allows them to do it, but this is an actual viable alternative,” Cofano said. “There are consumers buying and selling just with the help of AI now — it’s still on the edges, but it could eventually become mainstream.” 

For agents, Alloy Advisors said this means they will have to work hard to hone the skills that support tasks AI is not suited to do, such as mentally supporting consumers through what can be an emotionally charged process and negotiating on behalf of their clients. 

“There is a reason consumers dislike buying cars and it is because, for most consumers, the stress around negotiating on their own behalf is something that’s very unpleasant and they’re not good at it,” Cofano said. “So, it is important to highlight here that we believe that the role of the agent as active negotiator and trust companion in the transaction plays a meaningful role and is worth money. And I don’t think that is going away anytime soon.” 

Alloy Advisors hopes this serves as a wake up call for agents and brokers to begin thinking about what their business will look like in an AI-forward world. 

Regardless of exactly what that world will look like, Cofano and Kulkarni agree that the “great agent will win.”

“Great agents win because they put the consumer at the heart of their business. They generally do what is right for the person that they are actually serving and who is paying money,” Kulkarni said. “For brokers this means that many may need to pivot their model to put the consumer at the center and find ways to help the agent better serve this new well-informed consumer.”

This article was written by Brooklee Han and generated with the assistance of HousingWire Automation, then reviewed by a HousingWire editor before publication.