Ai Agent Platforms Look To Find Their Stride In 2025

Nearly one-quarter of all new startups funded by venture capital firms are artificial intelligence-driven companies, according to a PitchBook analysis published late last year. Given the massive changes experienced by the real estate industry in the past six months, it is no surprise that at least a handful of the startups feel that the real estate space is ripe for innovation, giving way to a proliferation of AI agent platforms.

As industry professionals and consumers have begun to settle into the business practice changes mandated by the National Association of Realtors‘ (NAR) nationwide commission lawsuit settlement agreement, HousingWire checked in with some of these AI agent companies to see how things are shaking out.

reAlpha

When it comes to AI agent startups, reAlpha is a bit of an exception, as it was founded in 2020 and became a publicly traded company in October 2023.

Initially, the platform used AI to help investors decide whether a property would be a good short-term rental property through platforms like Airbnb, and it acquired roughly 20 properties on its own. But when interest rates began to rise in 2022, reAlpha decided to pivot and eventually launched its AI agent platform that is 100% commission free for buyers.

Mike Logozzo, reAlpha’s president and chief operating officer, is still optimistic about his company’s future. But he is surprised at how little of an impact the business practice changes seem to be having on the real estate industry.

“We were under the impression that once the NAR changes took effect on Aug. 17, that there would have been a little bit more of an immediate change,” Logozzo said. “Historically, commissions were split between the seller’s agent and the buy-side agent, and we thought once the changes took effect, that would no longer happen. But that isn’t the case.”

Despite things not playing out as expected, Logozzo sees immense opportunity for his company’s offering.

“Our services cost buyers 0% of their purchase price, so that should be incentive for them to come to us, because now they can take the money a seller may be offering as a buyer’s agent commission, and since we own a mortgage originator, they can use that to help with their down payment or even buy down their mortgage interest rate,” Logozzo said.

Although it is still early, Logozzo said consumer feedback on the reAlpha platform has been positive. He attributes at least some of that to the fact that the company still works with human real estate agents.

“Our rationale is that we don’t want to leverage AI for everything,” Logozzo said. “We only use it as a tool to make it more efficient for our human agents.”  

Even with this outlook, Logozzo said reAlpha’s AI does roughly 80% of the work on a transaction, but a human agent always reviews an offer before it is submitted. Still, the brokerage still has some buyers who want to regularly engage with the firm’s human agents, while others almost exclusively engage with the AI.

The company is still only operational in Florida, but Logozzo said reAlpha hopes to expand nationwide later this year. In addition to the national expansion, reAlpha plans to continue looking for ways that AI can improve agent efficiency without replacing human agents.

Modern Realty

Launched in the spring of 2024, Modern Realty is one of the newer entrants into the space, but CEO and co-founder Raffi Isanians feels that his Y Combinator-supported company is on solid footing.

In the past few months, Isanians said, Modern Realty has become increasingly focused on creating the “ultimate real estate agent” by using AI as frequently as it can.

“Our reliance on solely AI at this point is not what we are doing,” Isanians said. “We are augmenting real estate agents up to the point where we can actually transition into being fully AI.”

This means that whenever a client asks the Modern Realty AI a question, or asks it to do something, there is a “very rigorous back-end process” in which a human agent reviews the responses to make sure the AI is not giving consumers bad advice.

“If something goes out to a consumer that is a mistake, it impacts our credibility, and when our goal is to become the ultimate real estate agent, that doesn’t help,” Isanians said. “This enables us to build a higher-quality product.”

Thus far, the company has found that most consumers are fine communicating solely with an AI agent up until the point when they are ready to make an offer. “People want to talk to someone when they are determining what number to put,” Isanians said.

Looking ahead, as Modern Realty looks to expand beyond the San Francisco Bay Area, Isanians said the company hopes to have just a handful of agents at the company whose sole focus is to negotiate deals.

“They would be a master negotiator,” he said. “The concept is that a traditional agent many only negotiate deals five or six times a year, but our agents do this five times before they even have breakfast.”

Homa

Another new entry into the space, Homa is not a real estate brokerage like Modern Realty or reAlpha. Rather, it is a tool that can be utilized by homebuyers whether or not they have a buyer agent representative. Despite this, Homa does not believe any AI product will ever fully replace a human agent for a consumer who wants to use one.

“I don’t think it is possible for the AI agent to be the actual agent on the transaction,” Homa co-founder Arman Javaherian said. “The best that we are going to be able to do is to have the AI be an assistant to those homebuyers who are trying to navigate this on their own.”

At the moment, Homa is free for users, but Javaherian said the company is exploring a variety of revenue models. Potential paths include charging users a small success fee for closing a transaction, or receiving referral fees from mortgage, title and home insurance partners.

The company is currently only operational in Florida, but Javaherian said he is hoping to see Homa expand to Texas in the near future.

Bramble

Also launched in the wake of the implementation of the business practice changes, Bramble was first conceived as an AI agent. But after going through Y Combinator and gaining more knowledge and experience in the world of real estate, co-founder Deepan Mehta said the firm has decided to take a bit of a pivot and, at least for the time being, focus solely on helping consumers find homes.

“We think there is a real opportunity there, because Zillow, Redfin and all of these other listing portal sites haven’t really changed anything in, like, 20 years,” Mehta said. “A really good Realtor is able to find out the exact specifics a consumer is looking for — those intangible things — and find their client properties that match their exact parameters. We think we can give people the same experience in their house hunt by using some of these new technologies.”

A licensed brokerage, Bramble supplies its listing platform via internet data exchange (IDX) feeds from MLSs. When a homebuyer begins their home search on the Bramble platform, its AI learns the buyer’s exact needs and wants, like a human agent would, enabling it to feed the best homes to an individual buyer.

“We want to offer people a super personalized experience, which an agent can’t always offer everyone,” Mehta said.

In addition to creating a more personalized collection of available listings for buyers, Mehta — who became a licensed agent — said that Bramble hopes to enhance listings by including things like video walk-throughs. Right now, the firm is only working with clients in the San Francisco Bay Area, but it plans to expand to other parts of the country.

“The goal is to make this really awesome platform where you go and you immediately start seeing really cool homes you are interested in,” Mehta said.