Marc Lore doesn’t pick his own meals most of the time. He has artificial intelligence for that.
And, he said, the AI is better at choosing his next meal than he is. “85% to 90% of my meals are AI derived,” Lore said on Wednesday at the North American Association of Food Equipment Manufacturers (NAFEM) Show. “AI tells me what I want to eat. I set my health goals. It knows what I like better than me.”
Lore is the CEO of Wonder, which has raised $1.6 billion and spent several years developing what it calls a “superapp” for meals. The company acquired the delivery provider Grubhub as well as the meal kit provider Blue Apron, which was recently launched on the Wonder app.
He has more plans for that, too. The company is looking at reservations and deals with grocers, all with the idea of giving people an array of options for meals that they can get delivered. “We want to service every meal occasion,” Lore said.
The AI program would take that even further, by allowing artificial intelligence to choose what customers should eat.
It’s not as crazy as you may think. People are using AI for a growing number of menial tasks, from help writing term papers to labor scheduling in restaurants. But it holds considerable promise and it’s not a stretch to imagine people using it to help with meal planning.
AI could take stress out of meal decisions based on what people like and don’t like. And it could in theory help people accomplish certain goals.
The AI platform brings together food and health, so it can suggest meals that help users solve for their health goals. For instance, if Lore needs to cut back on cholesterol, it may recommend some walnuts with the morning oatmeal. The AI program can’t make the food too healthy, because it wants to ensure a positive food rating.
A future iteration could solve other problems, Lore said. For instance, AI could help a user meet certain budget goals by sending certain groceries to your house. It then suggests meals so they get a fuller use of those groceries so there’s less waste.
“In the future people are going to give up a little bit of agency to solve goals,” Lore said. “AI is you, a smart version of yourself who is not persuaded by marketing or branding. It’s all based on science.”
(Check out our in-depth look at Wonder here.)
AI might be able to make a reservation at a restaurant you like and recommend what to eat off the menu based on what you’re craving or what you may like. But you can also tell AI to pick something else, which only serves to improve the program’s function. “It captures all the data,” Lore said.
As for Wonder’s restaurant-ghost kitchen-food hall hybrid, the company has 40 brick-and-mortar locations but expects 100 in 10 months and is opening one per week. It plans to open in Philadelphia and Washington, D.C., and has plans for places like Boston.
About half of its orders are pickup, and the concept thus far works best in suburban areas, though it is primed to go just about anywhere.
Its locations have a tight delivery radius to ensure food is delivered fresh and in less than 30 minutes. Orders have a 99% accuracy. Lore also said the company can operate as many as 30 concepts—each of which is its own brand—in one of its facilities with relatively low labor. Its locations range from 1,200 square feet to 2,800 square feet and can have 12 to 30 different concepts.
“We have one kitchen with lightly skilled and lightly trained labor, little waste, super high efficiency and the food quality is incredible,” he said. “We’ve spent hundreds of millions on culinary food science the past six years.”
Lore perhaps somewhat controversially argues that Wonder can operate all 30 concepts during off-peak hours with just two people. And he says that the company’s vertically-integrated business model means it doesn’t pay royalties to restaurants or to delivery providers.
“If you have a ghost kitchen, you have to pay royalties to restaurants, 30% to DoorDash and you have high labor and high waste,” he said. “You can see how the math is challenging.”
The long-term goal, he said, is to automate everything. “The goal is to move toward full automation in the future,” Lore said. “The food is robotically picked out of cold storage. The robot then takes it to the appropriate oven. That could pull a lot of labor out.”
Members help make our journalism possible. Become a Restaurant Business member today and unlock exclusive benefits, including unlimited access to all of our content. Sign up here.