When homebuyers begin their search, they’re not necessarily thinking most about price, size and bedroom count, said Creed Smith, a Denver-area broker and real estate technologist.
Instead, Smith said they’re picturing experiences — raising a family, dinner parties, Christmas morning and reading on the couch.
It’s exactly these dreamy aspects of homebuying that Smith wants to tap into with his new online search tool Envision, which lets users search for homes that match the style they like — irrespective of price, size, room count or architectural category. The tool is based on technology that outperformed a human broker in an Inman contest last year.
Though confined to Denver for now, Smith is marketing the technology at a cost of $29 per agent or brokerage website and $2 per agent for multiple listing services.
Envision users start by selecting the home style they like most from a gallery of preset pictures, then activate one of two search bots: “Dante” hunts for other listings that closely map to the selected style the user first chose, while “Eden” offers a more expansive set of options, including properties with related or similar styles.
For example, if a user initially clicks a photo of a bungalow, they will see local bungalows of varying prices and sizes at the top of the listing search results.
But as they dig deeper into the results, they’ll come upon homes that aren’t technically bungalows, but share qualities typically found in that style — such as a covered porch, tapered columns, hardwood floors and box-beam ceilings.
“Envision presents homes ranked by the qualities you most desire, allowing you to instantly see that $535,000 might get you a home you love as much as a $775,00 home,” Smith said.
Envision can detect more than 1,000 home characteristics based on listing copy, thanks to a years-long project from Smith to sequence the “DNA” of home styles. It can even predict a home’s characteristics from the presence of other attributes, allowing it to find inventory that wouldn’t normally appear in searches filtered simply by architecture, Smith said.
Envision also lets users turn on more conventional search filters such as bedroom and bathroom counts, square footage and price. And if a registered Envision user saves a property, they’ll receive an email notification every time a home with similar characteristics hits the market.
In a contest hosted by Inman last year, Envision’s bot technology beat human brokers at finding properties a real estate expert liked most, based on his selection of just one listing.
“I couldn’t really differentiate between what the bots picked and the humans picked,” the expert, John Rebchook, former real estate editor of The Rocky Mountain News, said at the time. “I tried to find a pattern to the picks, but could not. I guess if computers can beat us at chess and Go, they can best us at picking houses, too.”
Smith said that Envision represents breakthrough tech that can finally offer brokerages a competitive edge over the listing portals that dominate real estate search traffic these days. “This is their first chance in a decade to have better technology than the portals,” he said.
But large real estate brokerages and franchisors who expressed initial interest in Envision have thus far decided to forgo it, Smith concedes.
Smith said he wasn’t sure why, but speculated that part of the reason could be that Envision might appear to diminish the value of agents.
However, Smith said he was still optimistic that individual agents and brokerages would see value in his creation and its potential to attract clients and lessen their workloads. Industry heavyweights “don’t want technology to take over, but the really good agents do because they can see the advantage that it creates for them over their competition,” Smith said.
Some other property search platforms offer similar image-initiated search experiences. Trulia teases some property feature searches on its homepage, such as “homes with pools.” And RealScout, a platform agents use to conduct joint searches with clients, can compare listing photos on a room-by-room basis.
Smith said no other search tools use the same kind of exclusively image-initiated, style-and-characteristic matching method employed by Envision.