News

Better Service at Restaurants — With AI and Deep Learning

Coming soon to a restaurant near you: Artificial Intelligence.

AI is making appearances these days in many industries. 

AI is actively changing healthcare and finance every day. However, AI is still an infant in the foodservice industry. In 2016, food service accounted for 2.1 percent of the U.S. GDP. While that’s not massive, restaurants are the second largest employer in the U.S., and people interact with restaurants nearly every day. That means AI will soon find its way into your eating experience. Here’s how:

AI in Restaurants 

Robots will not be taking your order at the fine dining establishment down the street just yet. AI’s initial inroads in food service are an easy one: online ordering.

Chatbots are already working here, offering assistance as you place an online order for pizza. Chatbots and AI-based online ordering systems may fully automate restaurant takeout and delivery. That means fewer human errors in your food order.

Here’s how else AI will aid food service:

  • Recommendations. 
    — Netflix, Spotify, and Amazon are already telling you what you want. The natural next step is to have AI recognize our food preferences, even allergies or dietary restrictions. Food ordering apps can remind you of that dish you loved last time (or the one you hated) and recommend trying similar items if you’re nervous about something new.
  • Assessing food quality. 
    — Video intelligence can determine the quality of food to make sure things are fresh and have a good appearance.
  • Kiosks
    — Self-ordering kiosks are already appearing in some counter-service and fast food restaurants, including Panera and McDonald’s. These kiosks cut down on errors because people are putting in their own order instead of first telling someone else. Kiosks can save money for management and make the overall customer experience smoother, so you can expect to see more of these popping up in casual dining places.
  • Robots
    — While not appearing en masse, robots are helping in food service. A more common use is with food preparation or with delivery and optimization in warehouses and distribution centers. However, “Flippy” is helping cook burgers at CaliBurger and an actual robot waiter.

    “Pepper” takes your order at the MasterCard cafe and other locations in Asia, including Pizza Hut.

     

    (One company proposes a robot in consumer kitchens to prepare food and clean, but it may be awhile before you see it in your neighbor’s kitchen!)

AI in Restaurant Management

One of AI’s most valuable uses is data analysis. So far, there’s little mention of companies using AI to predict food orders, revenue, inventory, or customer traffic, but a few companies are starting to offer that data to restaurant groups.

One such program, Ingest.AI, takes volumes of data about food deliveries, shift hours, staffing, reservations, vendors, and bill paying. Most restaurants use a variety of software programs to manage related tasks, but none interact or generate a big picture, resulting in both food waste and wasted time and money.

This AI program “stitches all the data bits together and uses the data amalgam to give the restaurateur a 360-degree view of its operations, increasing the restaurant’s overall efficiency,” according to Wired Magazine.

Data can help restaurants automate scheduling.

 More than just scheduling is the forecast of busy or slow nights. Menus design based on popular items, manage inventory, and track staff who are better at up-selling or those give away too many freebies.

A restaurant might even know your preferences, making it easy for your server to offer your favorite dish and store your credit card on file for a hassle-free experience.

Deep Learning in the Food Industry

Deep learning may have a place in our food intake as well. Deep learning is a subset of AI. Deep learning uses artificial neural networks (ANN), a way of computing that mimics the human brain — except we have billions of neurons, while a large ANN consists of thousands of neurons.

These neurons work in many layers, from which we get the “deep” part of deep learning. Each layer can add and compute new information.

A Deep Learning Project

A group of Harvard students used deep learning in a project for TripAdvisor to help those searching for restaurants. The program took images uploaded by restaurants and visitors and classified them into five categories: food, drinks, interior, exterior, and menu. The results were 87 percent accurate on average. 

Photos

A similar project aims to predict restaurant attributes by analyzing user photos using 200,00 images on Yelp.

The Future of AI and Eating

The costs associated with AI and deep learning are too high for single-owner restaurants to manage. But as AI becomes less expensive and begins to save people money and time, you can bet it’ll be serving you up a plate of food in one way or another very soon.

No comments

Jeremy Webb

Chief technologi.st & Adventurer about.me/jeremy.webb

Jeremy WebbBetter Service at Restaurants — With AI and Deep Learning

Related Posts