By Tracy Morin

Between COVID-related changes in Americans’ dining habits, nationwide labor shortages, and tech advancements that are evolving and improving every day, operators are demanding more sophisticated phone ordering systems than ever before. “Even before the pandemic, pizzerias and other carryout/delivery phone order-driven restaurants were rapidly upgrading their phone systems,” reports John Scully, founder of PizzaCloud, a phone systems provider in Columbus, Ohio. “And, of course, during the pandemic, the need for a flexible, full-featured phone system jumped way up. Now, with ongoing labor issues, owners need even more from their systems.”

“When it comes to the ability to upsell and improve the guest experience, human conversation is still an important piece. But conversational AI has gotten smarter, and it can do certain things well.”
— Ray Pasquale, Unified Office


The AI Revolution

As technology continues to develop, many are looking to artificial intelligence (AI) to meet those needs. In a world where consumers are increasingly accustomed to interacting with chatbots and automated answering services to handle their inquiries, it would seem like a natural progression. “Five years ago, there would have been a very small number of people willing to interact with AI to place a food order,” Scully says. “Today, a much higher percentage are willing to do that.”

But what is AI, anyway? It’s a term we throw around so much that many people have lost track of its meaning (if they ever knew to begin with). Simply put, AI refers to intelligent machines that solve problems and perform tasks for us. For phone systems, it often means taking flesh-and-blood employees out of the equation. “Everyone is looking at AI, also called digital assistant, virtual call center, or, as I say, ‘the robot overlords selling us pizza,’” Scully says. “But what does that really mean? The best analogy is very simple: Think of it as a high-end third-party web ordering system, but instead of hands on a keyboard, the customer is speaking with it. Just as web ordering steps the customer through adding items, specifying options, upselling and possibly taking payment, the AI platform does the same thing, then pushes the order into the POS system.”

Related: How Erica Barrett is building the quick-service pizza kitchen of tomorrow


AI’s Limitations

Before you go leaping onto the bot bandwagon, keep in mind that AI still has plenty of limitations—at least for right now. In the first place, many people do still want to interact with human beings, and many operators prefer this option, too—if for no other reason than ensuring accuracy. “When it comes to the ability to upsell and improve the guest experience, human conversation is still an important piece,” explains Ray Pasquale, CEO and founder of Unified Office, a phone systems provider in Nashua, New Hampshire. “But conversational AI has gotten smarter, and it can do certain things well. For example, it can pull up a customer’s previous order and say: ‘I see you ordered X last time. Would you like to reorder?’” However, Pasquale stresses that you don’t want orders to be taken incorrectly or get lost altogether—and you definitely don’t want to lose a sale because a customer got frustrated with the AI’s automated “voice” and hung up.

Scully agrees—in his work with AI companies, he has seen a few common issues arise. Some, for example, “want to have something like ChatGPT read your menu and take orders,” he says. “Can it interact with callers? Yes. Will it make a ton of mistakes, resulting in bad orders, annoyed customers and furious owners? Yes.”

Other pitfalls he mentions include companies who do a custom build for each chain, sometimes investing 300-plus man-hours into creating the AI’s set of rules. “This tends to work well, but these companies usually have 25 to 50 locations,” Scully says. “You have five restaurants and want to use their AI? Nope!” Finally, he has seen companies that want too much money—up to $2 per order or 7% of the order (which can add up to far more). Scully points out that you can hire a human call center, with U.S.-based agents, for $2 per order. If AI isn’t a money saver, it’s difficult to recognize its value.

Today’s consumers are increasingly accustomed to interacting with digital solutions like bots, making them more open to AI-based phone ordering at foodservice operations. Photo courtesy Getty Images.


Finding Solutions

Regardless of these challenges, operators can take smart advantage of AI for their phone systems. Pasquale’s company, for example, has used AI to create something called Sentiment Analysis, a way to help maximize revenue and monitor customer satisfaction. “Say you receive 1,000 to 2,000 phone calls a month and there’s a dispute about an order,” Pasquale explains. “You can log into a portal, pull down the Sentiment tab, and choose keywords like happiness or anger. It’s 80% accurate in finding the calls you’re looking for.” 

United Office keeps adding more capabilities, such as real-time analysis, which can help create a score for a particular employee during a shift by determining how satisfied customers sounded. “Of course, in 2021, you couldn’t find humans to work, so we implemented different techniques of ordering, bots being one of them,” Pasquale adds. “For example, we do have a rollover option, where if the phone keeps ringing with no answer, you get a bot. And we’ve added drive-up window strategies, to make that process a lot more efficient.”

Related: Pizza Hut AI scanner knows the right food for your mood

One of the hurdles that PizzaCloud is attempting to overcome involves natural language processing. This means that machines can interact with human callers (customers) in a more fluid way, one that more closely emulates a real human being. And, since orders are going into a POS system, that’s another factor that needs to be considered when adding AI phone capabilities. Scully investigated 10 different companies to locate the best AI natural language processor company available, and is partnering and tightly integrating with POS companies that want to bring AI to their customers.

“In many cases, the POS companies are enhancing their Application Programming Interface (API)/data connections to give us the information we need to handle a higher percentage of calls well, and they are working with us to allow us to use automation to pull not just the menu but the ‘business rules,’ coupons, specials and everything else needed for the virtual call center to not make mistakes,” Scully says. “Would you use an inexpensive third-party web ordering system if one out of 10 orders was a disaster? No. Cost is important, but quality and accuracy are critical.”  

Tracy Morin is PMQ’s senior copy editor.

More Capabilities

John Scully, founder of PizzaCloud in Columbus, Ohio, notes that in today’s labor market, it’s a rare restaurant that has plenty of staff on hand. Thus, a phone system’s main task is often to reduce the number of calls ringing into stores. “In the first year of COVID, people wanted call queuing, auto-answering, and the ability to say things like, ‘Thank you for calling Super Pizza. If you have arrived for curbside pickup, press one. To place an order, press two. To hear what we are doing to keep you safe, press three,’ etc. So pressing one might ring only to a cordless phone carried by the runner, while pressing two rings the phones at the counter.”

Then there are other simplifying solutions, such as text-enabling the phone system so your greeting can say, “To receive a text message with links to our online ordering, press one. To speak to staff, press two.” “That way, callers who press one get a message with links to online ordering, or to download your app,” Scully explains. “The goal is to push people to order online, and it is very effective. I’ve seen locations move 10% to 20% of callers from phone to web orders in three to four months.”

Finally, there are options like virtual call centers, which can handle up to 90% of inbound calls with a heavy focus on upselling—so just as you see a higher average ticket on web orders than on phone orders taken in-store, you can see a higher average ticket using AI. This is useful for things like customers repeating a past order, checking on order status, or checking on pickup and delivery wait times. “None of those are show-stoppers if they’re not available,” Scully admits. “But they make a better experience for the customer—and fewer calls sent into the store.”