What is the worst possible time to launch a startup focused on the real estate market? Last year you might have been forgiven for guessing “In the middle of a global pandemic”. But as four ambitious university students are discovering, opportunity comes in many guises.
The real estate market in the United States took almost everyone by surprise over the past year. Although both the supply of, and demand for, homes plummeted in the second quarter of 2020, the recovery was swift — fueled by an exodus from Covid-19 hotspots, particularly cities, and a growing realization that working from home can offer remote workers far more flexibility in their location. Demand recovered, but supply remained historically tight.
With interest rates at unprecedented lows as well, homes began to rise in value rapidly. The U.S. Census Bureau tracked an average price of $374,500 in Q2 of 2020, rising to $434,200 a year later — a 16% increase that shows few signs of slowing down.
It’s into this red-hot market that the first-time entrepreneurs behind FlipAI are wading, as they build out an artificial intelligence-based system for non-institutional real estate investors.
“We see a lot of opportunity for more data usage to make markets more efficient,” says Srdjan Bozin, one of the four co-founders of the company, who is currently studying Computer Science at Stony Brook University. “Using data we are able to make it easier for sellers to arrive at the right price for their house, and for buyers to find the deal they’re looking for.” Fellow co-founder and student Prakash Sekar notes that the customer may be a “less experienced home buyer” seeking a passive stream of income from a rental property, for example or — as the company name implies — someone looking for an opportunity to renovate and flip a property.
FlipAI is built on the premise that smaller real estate investors — not the giant funds and private equity firms that are buying up entire neighborhoods — have to evaluate each potential deal almost from scratch, ascertaining not only the condition of the home but also its location, its proximity to schools and services, its access to broadband and so on. While each of these factors contributes to the value of the investment, there is currently no way for a small investor to compare and contrast potential deals without putting in an enormous amount of research.
Artificial intelligence powered by crowdsourced information is the solution, suggests Yee Aung of FlipAI, a junior at Harvard College. “An accurate model requires a wealth of reliable data across many parameters, a challenge that all real estate AI-powered analytics platforms face,” he explains. “We plan to add a wide range of parameters to the model sourced from publicly available data, including attractiveness of location, local economic data, seasonality, and so on. The synthesis of these parameters will be key to the model’s accuracy.”
Despite their youth and relative inexperience — the team met in High School, and their aspirations toward real estate investing are currently limited by available capital — the co-founders clearly recognize the inherent difficulties involved in building a business that relies on network effects for success. And public data sources alone won’t be enough, as Bozin readily acknowledges. “We need all this data before we have a useful product. We can’t just have a little bit of data and provide a little bit of analysis.”
Which is why FlipAI has launched on Multiverse — a startup platform for founders to test and iterate on their business plans, and which has a particular focus on AI-driven concepts.
Within Multiverse, individual startup ventures, or ‘planets’, are funded and supported at an early stage of development by enthusiasts. “Citizens working on our planet will have the opportunity to mark up images of properties and answer questions integral to assessing the valuation of a property, such as condition, design style, and more. Crowd-sourced data from citizens will allow our AI to account for local differences,” continues Aung.
To incentivize the crowd, rewards come in the form of FLIP tokens, a blockchain-based virtual currency that can be traded between the various planets in the Multiverse ecosystem. Aung notes that “Tokens can be held to obtain enhanced property intelligence and receive higher-paying tasks that require more trust and skill. The absolute supply of tokens, 76.74 million, is a shortened symbolic number representing the world population — participants in the housing market.”
Training the AI that will power the platform is a key focus of the founders, who are already seeking access to alternative data sources such as Matterport, a relatively recent innovation that maps the interior of homes and which provides would-be buyers with an opportunity to conduct virtual walk-throughs. “It has to be very objective,” says Bozin. “It can’t be something like, how would you rate this house out of ten. It’s the precise characteristics of the house, key information like ‘What kind of flooring is used in this room?’”
Using this data, the team intends to create algorithms that will reduce the friction involved in researching investment properties for buyers, and that will allow sellers to focus their attention on the type of upgrades or renovations that secure the best possible return on investment.
The fourth member of the team, Daniel Verbesey, studies Applied Mathematics at the University of Pittsburgh. On monetizing the business, he explains that “We’re still deciding between a subscription model, where customers can pay monthly or yearly for select data, or a model where customers purchase a specific set of data — like a local dataset. We’re also considering how sellers can pay to get an accurate estimate on their home.”
As youthful entrepreneurs, none of the co-founders seem fazed by their relative lack of exposure to real-world businesses. Instead they are eager to gain experience rapidly. Aung, for example, has previously worked on a startup incubated by the Harvard Innovation Lab and has an internship at Roche, while Verbesey is Vice President of the Pitt Blockchain Club and both Sekar and Bozin are interning as software engineers at a sports analytics business.
“I have a friend in college who started a Bitcoin ATM company,” says Verbesey. “Within a few years they had 13 different locations. If you see the opportunity and don’t go for it, someone else will come along… so why not just build it yourself?”
Bozin believes that a startup founder’s youth is less important than “If you’re actually solving a problem… If you really take the time to talk to your customers and see what issues they have and where you can help solve them.” To that end, the team is engaging frequently with the Multiverse community, which Aung describes as “changing the centralized landscape of A.I. for the better with decentralization, and also equipping like-minded innovators with a platform to collaborate.”
“We’re aggregating the opinions of mentors to see if our idea is going in the right direction,” explains Sekar. “Maybe it’s a pitfall for young people, that they become too emotionally-invested in an idea, and foresee a future that might not really be there.” But he feels that testing the concept in a startup ‘sandbox’ has helped the team behind FlipAI understand their product-market fit more deeply.
As FlipAI matures, their data becomes more reliable, and specialized algorithms begin to deliver real-world value to investors, Aung hopes that the company and the Multiverse platform on which it’s built will continue to “foster a community of citizens whose motivations are aligned with our planet.”
And even though the real estate market will eventually cool off, perhaps before their platform is ready for mass usage, these four emerging entrepreneurs will have plenty of time to refine their tools before the next big boom.