Proptech Rides to the Fight Against AI-Driven Rental and Mortgage Fraud

It doesn’t take a Columbo or a Holmes to uncover financial fraud in real estate, because artificial intelligence is on the case.

As rental and mortgage fraud is a growing problem — in part due to AI — owners and proptech entrepreneurs are increasingly applying the technology to stem the tide. 

Overall, U.S. consumers reported more than $10 billion in losses due to fraud in 2023, according to the Federal Trade Commission (FTC), with identity fraud reaching over $40 billion, per Javelin Strategy & Research, an independent financial services research and insights firm.

Property managers and landlords are suffering a lot of fraud-related losses, too. In 2025, more than 12,000 complaints related to real estate fraud were filed with the FBI, totaling more than $275 million in reported losses, according to bureau data. Since 2020, $65 million has been lost in rental scams alone, a December 2025 FTC report noted.

Mortgage fraud is also growing, with an estimated 1 in 118 mortgage applications having indications of scams, according to a November 2025 report from Cotality, a property information, analytics and data-enabled solutions provider. Its National Mortgage Application Fraud Risk Index for the third quarter of 2025 found mortgage fraud risk increased by 8.2 percent year-over-year.

Vero, a multifamily leasing platform that replaces manual renter screening with real-time identity and income verification, is a company attempting to address rental fraud. The Manhattan-based company raised $810,000 in May 2025, led by Sunriver Capital and Fifth Wall, bringing its funding total to $19 million, according to Crunchbase.

“Vero was one of a few first movers in what I would consider to be Version 2.0 of screening in multifamily,” said Copley Broer, chairman of the board of Vero, as well as managing partner and co-founder of Sunriver Capital. “Historically, screening was relegated to credit criminal eviction, and those are fairly commoditized by businesses such as TransUnion, Experian and people like that. Vero was really early on into the next phase, which was ID verification, income verification and more fraud-related documents. So we felt it was a very good technology platform and a good place to build off of.”

The exponential growth of artificial intelligence has quickly increased what Broer called “the front end of fraud,” the aforementioned ID verification, income and document fraud.

“AI has made it really easy to tamper with stuff,” he said. “So it’s now not that big a deal for somebody to change the amount of money they make on their W2 and upload it, or to change the name on what looks like an otherwise valid driver’s license. AI has made that super simple for people to commit that sort of fraud.”

This has resulted in leasing teams often being expected to spot fraudulent documents with little to no formal training, while AI-generated financial documents are making traditional screening methods increasingly unreliable. Additionally, the industry is shifting from document-based verification to direct-source data, including payroll and banking data. The AI-driven automation velocity is also increasing the tension between faster leasing decisions and rising bad debt and fraud risk, according to Vero.

Industry groups and credit bureaus are warning that generative AI is accelerating synthetic identity creation and documentation manipulation.

Some instances of real estate fraud still go overlooked, though, said Steve Carroll, co-founder and CEO of Findigs, an AI-native leasing platform that automates rental screening for property management companies and owner operators.

“The real estate business is a significant affair, with many people and extensive documentation involved,” Carroll said in an email. “This combination of factors leaves room for manipulation. One can say with certainty that the $275 million figure in reported losses is underestimated, as much of the rental fraud goes unreported due to property managers absorbing the loss quietly or not realizing what happened until well after an eviction.”

While the rental sector has long relied on credit score checks and one-time document verification to weed out fraud, generative AI is posing an increasing challenge for the industry, Carroll said.

Vero’s Broer noted that the rental industry has a tremendous number of applicants who have multiple incomes, with approximately 65 percent having extra income in addition to their standard W2 earnings.

“The ability for a quote-unquote underwriting or screening provider to accurately underwrite income has become increasingly difficult,” he said. “Vero works to put those technologies that are more difficult and less commoditized ahead of what you would consider to be traditional screening. It then takes all the data that it learns through that process of onboarding and underwriting and feeds it into our decision engine, which weeds out fraud at a higher pace than than other solutions.

“It is also able to make better decisions on how to get folks into buildings that do need to be there, that are not fraudulent, increasing leasing velocity as a result. So we try to toe a fine line between being a blocker of fraud, but also not inhibiting the occupancy rate or the rate at which these apartment owners are able to fill their units.”

Affordability is driving some of the growth in rental fraud, particularly in Class B buildings and in areas such as Atlanta, Houston and parts of Florida, said Broer. 

“We see an affordability mismatch in Florida that we don’t see in Texas,” he said. “Things like that tend to pile up in a specific geography because rents have gone up — especially in some major metropolitan areas at a really precipitous rate compared to wages. So there is just the fact that fewer people are able to afford the rent-to-income ratios on fewer units.”

The rental world is not alone in battling fraud.

Cyber scamming is such an issue in the mortgage industry that large mortgage originators have already started incorporating technology into their operations, said Dave Parker, CEO at LoanLogics, a Jacksonville-based provider of technology, software and automation for the industry. And, now, even smaller lenders are researching AI tools to help improve their operations as well as prevent fraud, he said.

“It’s always there,” Parker said of mortgage fraud. “And, during periods, there’s always upticks. If it’s not growing now, it will in various areas, so you have to be ever-vigilant to detect it.

“In regard to detection, a lot of that’s around triangulation of a lot of the data inputs. If you say you’re making a million dollars, but you’re working part time at a fast-food restaurant, then that’s not adding up.”

Founded in 2006, LoanLogics has more than 500 clients, including over 50 percent of the largest originators in the U.S. The company provides a product called LoanSphere, which helps determine monthly qualifying income for representative and warranty relief immunity from Freddie Mac and Fannie Mae repurchasing demands.

In addition, LoanLogics uses data triangulation and integrity checks to detect fraud, including altered PDFs and inconsistencies in loan data, from among 1.34 billion unique documents and 4.8 billion loan pages processed. It also employs a software called LoanHD to identify variances and flag risks.

LoanLogics is already seeing that the next wave of AI will automate many of the functions previously assigned to offshore businesses, leading to enhanced security and reduced operating expenses. But as AI adoption continues to accelerate, safeguarding the path and integrity of mortgage data will become just as important as the technology itself.

“I think it’s an issue,” Parker said of the potential increased use of AI to commit mortgage fraud. “When you go and ask AI to generate bank statements, for instance, it’s very smart, very savvy, and it can do that, which makes triangulation of data even more important. Data that you used to always be able to take from statements and treat it as evidence is now something that you have to cross-check and validate from other sources more and more.”

The current state of mortgage fraud defense has caused an AI race between companies like LoanLogics and scammers, said Parker.

“Without a doubt, it is a race,” he said. “And you have to understand all the flavors of AI, how you want to use it. There’s an incredible learning curve, and it changes in real time. You have to be very assertive to be able to keep up with it, and very aggressive at how you test it and determine the applicability of different types.

“We’re very cautious about AI, where the logic can drift on you, particularly when you’re making any type of decisions around lending, which is very rule-based, very black and white,” Parker added. “So, depending on your need, there’s different vintages, different tools of AI that you want to employ.”

Philip Russo can be reached at prusso@commercialobserver.com.