When people talk about artificial intelligence and mortgages — and they do, a sign of AI tackling more real-world problems — they mostly refer to making the process less of a paper-heavy slog.
Then there’s Proportunity.
Born of a unique tech startup builder called Entrepreneur First — a “talent investor” that pairs up smart strangers and helps them develop whatever company idea they come up with — Proportunity uses AI based on Intel technologies to find London properties that are poised to appreciate in value. The company then offers a sort of “bridge loan” for first-time buyers who can afford the monthly payments but don’t have enough for the down payment that traditional lenders require.
“We started with first-time buyers because this is a huge problem here in London,” says Stefan Boronea, Proportunity co-founder and CTO. “Affordability is pretty low.”
The company estimates that down payment requirements keep some 200,000 more would-be first-time buyers from joining the U.K. market today, compared with 2001.
Proportunity helps bridge that gap with an interest-only equity loan on top of a traditional mortgage, thus lowering the initial outlay required to complete the purchase. As the home’s value appreciates, the homeowner can refinance and pay off the Proportunity loan.
“Customers want to be confident that they’re buying a property that will grow in value, and we want to drive good investments both for the customer and for our investors,” says Boronea.
So how does Proportunity find diamonds in the real-estate rough — with so much confidence? The company uses a combination of spatiotemporal machine learning algorithms, Boronea says, “to understand what macroeconomic and microeconomic factors are driving price growth.”
The data is out there, Boronea explains, but the hard part is figuring out which data is relevant and then using it to build an accurate prediction model. “There are a lot of new potential data sources that we could use and new experiments to try,” he says, “but there’s always the question of how do we prioritize that and how do we test really quickly what works and what doesn’t work.”
In other words, time is precious — and that’s where Intel helped give Proportunity a boost. Its data scientists worked with engineers from the Intel AI Builder team, who shared expertise and introduced a set of Intel-optimized software libraries and tools into their workflow (read more about Proportunity’s technical solution).
After testing various CPU- and GPU-based systems, the company now relies on Intel Xeon Scalable processors for both its AI training and inference. “We have a pragmatic view on technology,” Boronea says, “and the training and the testing proved to be very successful for us.”
Proportunity recently celebrated its first year of issuing loans — meet its first customer on YouTube — and is working to expand. “There’s a huge need worldwide for this kind of product,” Boronea says.
“Globally, property markets respond in similar ways to fundamental shifts in population, age, crime, accessibility and of course, affordability,” Proportunity CEO Vadim Toader told City A.M. “Now we have the tools to process these shifts, predict the reaction, and get that information to those who need it.”
That means first-time buyers struggling with down payments in other cities might get an AI-powered helping hand soon, too.
下载 ChinaIT.com APP，随时掌握最新IT资讯