The Infatuation's Text Rex product provides personalized recommendations. / Skift Table The Infatuation's Text Rex product provides personalized recommendations. / Skift Table

Restaurant Megatrends 2018: Good Content Is the Path to Bot Recommendations

Download your copy of Skift Megatrends 2018

We’ve just released our first annual restaurant industry trends forecast, Skift Megatrends 2018. You can read about each of the trends on Skift Table as well as download a copy of our magazine here.

Scoring the absolute perfect restaurant recommendation — from a friend, from a stranger, and even from a brand — feels almost magical. As we turn to everyone from our friends to our phones for advice (and now even Amazon’s Alexa has opinions), the prevalence of text-based communication can sometimes make you wonder if there’s actually a human on the other end.

More often than not, there is. And until the robots can take over, all companies are smart to adopt strong content strategies to boost their cache of information about restaurants.

“What we often see today is a list of every restaurant we think is best,” Reserve CEO, Greg Hong, told Skift Table last September. “Everyone has their different version of what best is depending on the day of the week or who they’re going out with. How do we find the best restaurant for you then?” The answer of the future lies in technology, but the process of today depends on robust content and information about restaurants.

Personalization is a buzzy product feature for any tech-based company. Now that we’re conditioned to the world of on-demand everything, the next logical step becomes companies offering the exact right option at the exact right moment. And in an age of increased expectations around digital services and products, the stakes around getting it right have never been higher.

The Beauty of the Human Touch

In the early days of searching for a place on the internet (remember Citysearch?), sites input a ton of information for each restaurant or venue. Tags indicating types of cuisine, the restaurant’s vibe, hours and location all helped drill down which to recommend upon search. While these facts are clearly important to finding the right place to eat, they don’t tell the full story. Surfacing a bunch of Italian restaurants in your neighborhood that happen to be open is useful, but which are actually good? Which have dim, romantic lighting? Which has the best table for a first date?

Answering these particular questions is one reason restaurant review site The Infatuation launched its Text Rex service. Launched in New York in 2015, Text Rex is a messaging platform for restaurant recommendations: text your needs to a number, and get a fast, personalized, appropriate response. And, yes, there’s a human on the other end of the line. (Text Rex is in Los Angeles now, too, through a partnership with Delta.)

Infatuation co-founder and CEO Chris Stang admits that people — as opposed to bots or computers — handling the interaction is challenging, especially at scale. “We find running it with humans really is so much more effective at building a relationship and also helping people get what they want,” he told Skift Table.

Of course, these personalized recommendations are backed by a whole lot of content. The Infatuation operates restaurant review sites in nine cities (and publishes guides for many others), and has editorial reviews, rankings, and recommendations for hosts of restaurants around the world, all of which are available to the team of over 40 approved “Text Rexers.”

“The reality is that conversational AI is not in a place yet where it can handle complex conversations with the nuances involved in helping someone find a great restaurant,” said Stag, but it is getting there.

Robots Begin Pulling Their Weight

Currently the push toward personalization drives development of both AI and human-powered recommendations, and finding the right restaurant for the right moment is top of mind for most companies in the space. Reservations platforms like OpenTable, delivery platforms like Grubhub, and review and recommendation sites like Yelp all surface recommendations through editorial curation, artificial intelligence or machine learning, or some combination of both.

Artificial intelligence in the form of chatbots already works well for plenty of restaurants. Chains from TGI Fridays to Shake Shack use bots to answer menu questions or give directions to a location. Some even accept orders via Facebook Messenger or through the restaurant’s website. This is a great boon for businesses, but the technology isn’t quite ready to replace humans at scale. The answer to the complicated puzzle of providing personalized recommendations may lie in this smart technology, but the secret to success right now lies within all of the underlying content that powers the technology.

What to Expect

Use of this type of tech is still in it’s early days, but there’s good news for the future given the underlying editorial curation as it continues to grow. “People are more and more familiar with the idea of having a conversational interaction with a brand,” said Stang.

On the path to robot takeover, companies are building ways to determine preferences and build recommendations across the board. “Personalization as a concept is something we’ve been doing for a while,” said Uber Eats project manager Ambika Krishnamachar. “Everything is personalized based on the taste profile we infer about you.” The Uber Eats taste profile is compiled using information about the restaurants you order from and browse on Uber Eats. Krishnamachar compares the predictive technology to that from Netflix or Spotify, suggesting restaurants or particular dishes it thinks you want.

Even with a large team and the tech powerhouse at Uber behind them, the computers don’t do all the heavy lifting — human touch still factors into play. Uber Eats reviews, for example, measure current sentiment about the restaurant by creating a rating based on the last 90 days of review data about a particular restaurant.

Although Uber will continue to invest in its personalization technology, editorial curation — and human input — won’t go away from Uber Eats’ product offering. “We have a lot of personalization technology to help learn the user preferences,” said Krishnamachar. “At the same time, there’s a particular art to the way that restaurants construct their menus. I think we’re still kind of looking for the delicate balance for how the two will interact in the app.” She imagines the “recommended” feature in the app might evolve into a space where both AI and input from the restaurants determines what a diner sees.

And even given The Infatuation’s commitment to the human touch, it’s also testing a mix of human-AI interaction “to help a conversation get to the point that it can be passed off to a human,” according to Stang. Of course, he added, the smart software isn’t the only secret.

“Also, magic.”

Download Your Copy of Skift Megatrends 2018

This year’s Megatrends are sponsored by our partners at AccorHotels, Allianz Worldwide Partners, Hilton Garden Inn, Intrepid Travel, onefinestay, and Upside.

More from Skift Table