The Next Big Thing In Algorithm-Driven Recommendation Services? Humans

Text Rex, from the Infatuation food and drink site, gives local search a personal touch.

The Next Big Thing In Algorithm-Driven Recommendation Services? Humans
[Photo: Flickr user Georgie Pauwels]

I’m texting with a stranger about dive bars. No, this isn’t a Tinder date in the making. I’m just looking for places to go hang out after work. But rather than firing up Yelp or letting Foursquare push machine-generated ideas my way, I’m using a service that puts an old-school spin on local discovery: It lets you talk to human beings.


Text Rex is a new feature from The Infatuation, the New York-based food and drink review site. It’s not an app so much as a service that lets you text with a local expert, typically a staffer from The Infatuation. Using the site’s trove of reviews as a guidebook, coupled with their own knowledge of local spots, this small army of invisible tour guides points users toward the best dining spots near them, one on one.

“You can’t ask a website or an app follow-up questions about a restaurant if you need more context,” says Chris Stang, CEO and cofounder of The Infatuation. With Text Rex, he says, it’s different.

Indeed, it is. Using Text Rex feels conversational and natural, more like asking a friend for restaurant tips than pinging a database. That’s because The Infatuation has done something that’s oddly novel nowadays: It’s injected a human layer between the database and the query. In this age of automated, algorithmically driven everything, it’s surprising how refreshing that can be.

After signing up for Text Rex, I introduce myself and let them know that I’m in the East Village and in the mood for a diner type of vibe. (The truth is, I’m actually in Philadelphia, but this is an NYC-only service for now. I can always go for a milkshake, though.)

Within seconds, Text Rex recommends a place called La Bonboniere and sends me a link for more info. A minute later, the person corrects him or herself: “Ah, you said East Village. That’s West.” In a strange way, this screwup feels refreshingly human. Algorithms and data flub details all the time, but when it happens, it’s a dead end for the user. By confusing the East Village with the West Village, whoever’s on the receiving end of my request is exhibiting a very human trait. It’s not a letdown, either: To me, the whole thing feels more authentic this way. And isn’t authenticity in during the artisanal age?

Text Rex sent me this animated GIF of a scene from The Office to express approval. Take that, search engines!

Dive bars in Williamsburg. Lunch spots in Queens. Whatever you’re in the mood for, Text Rex has an answer. It also has a sense of humor. After I test the waters with an off-color joke about launching a similar service for buying marijuana, Text Rex sends me an animated GIF of a scene from The Office in approval. Try doing that with a search engine.


Whether or not the east/west gaffe was deliberate, the more human feel definitely is. “We believe strongly in the fact that human intelligence is crucial to this kind of service,” says Stang. After soft launching a few weeks ago, Stang says Text Rex has been well received, with many users commenting on how similar the service is to the role friends play in each other’s lives. In fact, many have asked how they can get involved and become recommenders on Text Rex.

That’s a good sign, because as you might imagine, a service that relies on human brain power and manual effort isn’t easy to scale. That’s their biggest challenge, Stang says. Over the next six months, the company hopes to expand the service to other markets that The Infatuation covers: Los Angeles, San Francisco, and Chicago. To keep Text Rex feeling like a well-informed friend, it needs to be both knowledgeable and fast. They can’t have an intern sitting there fielding hundreds of inquiries at once. But should Text Rex see a sudden spike in users, Stang assures me, they’ll be ready.

“We’ve got a mix of staff and contractors running it,” says Stang. “We have an equation worked out on how many users each human can handle with the technology included,” says Stang. “We feel good about that equation as it stands right now. We’re ready to handle a lot of new users.”

The Infatuation team has thrown around ideas about how to monetize the service–they refuse to make paid listings part of their business model–but Stang says revenue isn’t a major focus at the moment. Instead, they’re fixated on the user experience: how to onboard new users and get them in touch with a real, live human as quickly as possible. Somebody that talks like a friend, not a machine. You know, somebody who might confuse the West Village with the East Village.

After correcting the mistake, the person on the other end of my Text Rex chat offers up another suggestion: “How about Veselka?”

I’ve been there before. Good choice.


About the author

John Paul Titlow is a writer at Fast Company focused on music and technology, among other things.