2023 Vertical Markets Spotlight: Speech Technology in Restaurants

The restaurant industry has often been a laggard in adopting new technology. But with speech tech, that is not the case. The need to streamline business processes to maximize staff productivity and improve customer satisfaction is driving companies to implement speech technology for taking orders and servicing customers.

The industry is highly competitive, high-risk, and volatile. Profit margins are typically low, and staffing problems are common. Restaurants serve large numbers of customers in short time windows, with a mad rush of customers during breakfast, lunch, and dinner, when they have to produce meals and turn tables quickly and efficiently. Wait times in lobbies and at tables, counters, and drive-throughs lower revenue and increase dissatisfaction.

To say that the past few years have been tough on the restaurant industry would be an understatement. Even before the COVID-19 pandemic, the market was in flux. Restaurants saw dramatic increases in delivery and takeout business with the emergence of services like DoorDash, Grubhub, and Uber Eats. In 2019, 60 percent of U.S. consumers ordered delivery or takeout once a week, according to several reports.

The pandemic turned traditional business models upside down as restaurants had to completely end indoor dining. Takeout, which had been a growing part of their business, quickly became their only income source.

COVID also ate into the restaurant industry’s labor pool, and many of the people who left their restaurant jobs during the pandemic have not returned now that the worst is behind us.

“Four out of five restaurants are understaffed today,” says Ben Bellettini, vice president of restaurant sales at speech technology vendor SoundHound, which offers SoundHound for Restaurants, a system that can accept voice orders from customers over the phone, via menu kiosks, or at drive-throughs. “The average tenure for a new employee is about nine weeks. So restaurants spend time searching for staff, find them, train them, and then lose them.”

Many restaurants simply lack personnel to pick up the phone when customers call, so those people go somewhere else to eat. And when consumers do show up, they often face lengthy wait times and poor service.

Consumers Embrace Restaurant Chatbots

To address these problems, many restaurants are turning to voice solutions for help. Voice input offloads manual work to machines and eases the personnel crunch. It can take customers’ orders and route them to supporting food preparation, delivery, and payment systems. In best-case scenarios, artificial intelligence understands the words customers speak, correlates them to specific items on the menu, and routes orders directly to the kitchen.

Consumers seem to understand the benefits that speech recognition offers everyone involved in the process. In a 2019 survey, nearly two-thirds of American consumers said they would like to use voice-based systems to order food, and those numbers have only been rising.

In fact, 65 percent of customers said they would visit a restaurant more often if self-service kiosks were offered, and 50 percent said they would prefer to order at kiosks if lines are longer than four people, according to research from Tillster.

Voice technology could empower food service vendors and customers in so many ways, but it also creates hurdles. Designing, iterating, deploying, and maintaining voice solutions are not for the faint of heart.

Voice AI has been a complex, resource-intensive emerging technology, and most restaurants typically do not have the data scientists, data analysts, and database system designers needed to build and maintain the systems. Even when they do manage to incorporate these systems into their day-to-day operations, they face a learning curve and systems integration challenge.

Voice-based ordering assistants need to be integrated with restaurants’ computer infrastructure, including its point-of-sales terminals, web pages, mobile apps, kitchen workflows, and back-office solutions.

It becomes an expensive proposition. “Voice deployment costs are high because users pay a number of licensing fees,” notes Seth Sternberg, product manager at CEVA.

When food service providers do roll out voice systems, they face challenges with the technology itself. Drive-through windows, for example, are high-traffic, noisy areas that can greatly impede recognition accuracy. Understanding not only what a person said but what he meant can be quite difficult.

Most voice systems focus on a narrow group of commonly used phrases to derive meaning from customers’ requests. But individuals speak in unique ways, mixing and matching phrases. One person might present an order as a command: “Give me a hamburger.” Another might pose it as a question: “May I have a hamburger?” AI algorithms are hard-coded, follow set patterns, and can have trouble putting all of the iterations into the proper context.

Modern voice systems can handle many types of inputs, but sometimes they fail to recognize what a person says or needs. In those cases, it’s important for the eatery to have an employee available to step in and complete the transaction.

The Business Benefits

While implementing voice technology requires a lot of work, the effort pays off in many ways, starting with addressing personnel issues. “Restaurants have thin staffs and need to offload as much as they can to computers,” SoundHound’s Bellettini explains. “Placing voice self-service systems in the front of the house enables them to use personnel in the back end where the meals are prepared.”

Modern consumers demand instant gratification, and restaurant voice systems offer it. For starters, with orders processed more quickly, customer wait times become shorter, food gets to the grill more quickly, and customers are on their way with fewer delays.

Employees then have more time to connect with customers and provide more personable service.

More sophisticated voice assistants can also provide answers to customer questions or make ordering recommendations based on product availability, special offers and promotions, or customers’ stored order histories or dietary restrictions. That can increase customer satisfaction, loyalty, and even order value.

Automating tasks like taking orders or scheduling reservations can increase profitability in other ways, too. Ideally, orders going into the system are immune from human error, so less food is sent back to the kitchen.

With so much to gain, it’s not surprising that many food service operators are scrambling to move the technology from the development labs and into the field. The number of chains testing it has been rising, and the initial results have been promising.

Del Taco serves more than 3 million guests each week at about 600 restaurants across 16 states. In January the company started piloting Presto Voice from Presto Automation. An AI-powered voice assistant greets guests, accepts orders, and offers upsell suggestions. During that pilot, 95 percent of all drive-through orders placed using Presto Voice were completed without staff intervention and automatically routed to POS systems and Kitchen Display System applications. As a result, Del Taco now plans to roll out Presto Voice at additional locations nationwide.

McDonald’s introduced voice ordering at 10 of its Chicago drive-throughs. Users drive up to the screen and place and pay for their orders through a voice interface. The software—which flowed from the fast food giant’s purchase of Apprente, an AI voice recognition technology provider, in 2019 for an undisclosed amount—accurately recognized orders about 85 percent of the time. McDonald’s plans to use the system in more locations.

Good Times Burgers & Frozen Custard tested the Valyant AI for breakfast in a few of the 35 locations it operates in Colorado. The time between when customers placed their orders and drove away with their meals dropped 10 percent. The pilot’s success led Good Times to expand the offering to lunch and dinner.

Several other companies are also turning to AI-powered solutions to help with orders. Checkers & Rally’s are using drive-through technology from Presto and Valyant AI. Wendy’s is using Google Cloud’s AI tools to improve the guest experience, and Papa John’s recently added an AI-assisted call center, called PapaCall.

SoundHound for Restaurants is also seeing some impressive rollouts, with White Castle and US Café among them.

“Restaurant customers want to place food orders quickly and with minimum hassle. Our fast, accurate voice technology allows them to do that by speaking as naturally as they would to a person,” said James Hom, chief product officer of SoundHound, in a statement.

Restaurants have been struggling during the past few years with changes in how they deliver food amid a severe personnel shortage. Voice interfaces are helping them address those problems. But the integration work required is complex and challenging. A growing number are taking on the work, and the initial results have been promising.

Breaking with tradition, this market sector is at the forefront of voice recognition technology deployment, and more is coming. In fact, the growth of AI and automation in restaurants is likely to continue, with 50 percent of U.S. operators planning to implement some form of automation technology in the next two to three years, according to a Lightspeed report. x

Paul Korzeniowski is a freelance writer who specializes in technology issues. He has been covering speech technology issues for more than two decades, is based in Sudbury, Mass., and can be reached at paulkorzen@aol.com or on Twitter @PaulKorzeniowski.