Foodler Debuts Diner's Best Friend: Virtual Waiter
Foodler is ready to be the Alfred to a diner’s Bruce Wayne.
Think of butler Alfred being able to decisively tell Bruce Wayne which foods to order at a new restaurant. Batman’s appetite would be pretty satiated, right?
Christian Dumontet and John Jannotti, co-founders of food ordering website Foodler, initially set out to create a virtual waiter. In actuality, however, Jannotti says they one-upped the waiter.
“You might even call it a virtual butler,” Jannotti says. “A waiter doesn’t know you, but we know you. We know what you’d like to order.”
Using artificial intelligence and data analysis, the Foodler Recommendations Engine recreates the restaurant experience by suggesting new dishes based off the diner’s previous orders. It integrates flavor and taste preferences, as well as personal modifications, to siphon out undesirable cuisines and hone in on uniquely delectable dishes.
On the website, a user then sees a tag cloud highlighting the particular tastes or flavor profiles that he orders most often.
“Between the personalization and the ability to tell you new things about restaurants you haven’t even visited yet, we thought this was an easier way [to order],” Jannotti explains.
Restaurants benefit as well, Jannotti says, because they no longer have to guess which dishes to promote; Foodler can tell them with certainty.
“Restaurants try to put their most appealing dishes first or at the top of the category or under a category called House Specialties to draw attention to it,” Jannotti says. “But really, that’s just guesswork or the most popular dish in general. We’ll recommend actually the single dish.”
While the virtual waiter suits some users, Jannotti says there are still kinks to work out.
“We had a woman who was absolutely upset because she was a vegetarian and some of our recommendations had meat in them,” he says.
The system recognized that the client preferred fruits and vegetables, but she had not yet ordered enough from Foodler for meat to be eliminated from her recommendations.
“The reason that I like that story, even if it left us with something to improve, is because it mattered so much to her,” Jannotti says. “We were really happy to see that people are paying attention to the recommendations and they expect them to be good, and we're going to prove that we've already working a system out where we notice faster classes of food. We'll learn faster.”
By Sonya Chudgar