Think you’re busy? Consider the plight of QSR cashiers wearing headsets to answer calls while taking in-person orders and coordinating with the back of the house to expedite food. It’s not easy, and we’ve all been on the business side by calling in to place an order and immediately being put on hold. While this very situation has fueled the growth of third-party delivery, there’s a new option to improve this arrangement for workers while trimming labor costs, increasing upselling and ensuring that every caller ultimately gets to place an order. The secret sauce is automated call answering powered by artificial intelligence.

Adam Ahmad is the founder and CEO of Kea, which recently began offering computerized, cloud-based call ordering that eschews the long-standing issues with phone calls coming into busy restaurants—for employees, owners and customers alike. Previously the founder of an early third-party deliverer and later an employee with employee sharing firm Next Force Technology, Ahmad’s new thing is reviving the experience of calling into a restaurant, which many have written off for dead. That crew of skeptics may include a few restaurant journalists, as well.

Through his work at Next Force, Ahmad noticed how burdened QSR cashiers were, combined with the difficulty of attracting and retaining such employees. With the volume of calls that still come into restaurants, like a Pizza Hut, he wondered how much money such restaurants lost by placing most everyone on hold, missing some of those orders due to extended hold times, and loading down employees who have larger fish to fry, figuratively speaking.

“I sat on a bunch of panels and noticed a theme,” he said. “The COO at one of [those restaurants] was like, ‘Guys, we’re losing between 30 and 40 percent of revenue because we’re not able to pick up the phone’—that was the aha moment for me.” Later on in the presentation, improved scheduling and hiring more people was offered as a solution. Thinking about the nationwide challenges of restaurant staffing, Ahmad realized you couldn’t “throw a human at the problem.” Enter artificial intelligence.

He decided to conduct an experiment with a franchisee who was a former client, forwarding all of the restaurant’s inbound calls to his personal cell phone for a week. At the end of this chatty week, they compared sales to the previous week and found a 19 percent uptick just from his stint working the phones.

As the story often goes, he soon quit his job, dove into developing technology to automate customer calls, and embarked on an ultimately successful hunt for seed funding: and thus, Kea was up and running.

Smarter and more responsive than the now-familiar invisible friends like Siri or Alexa, Kea’s technology creates a true conversation where customers can ask questions—like what are today’s specials—and the artificial intelligence always remembers to suggest add-ons like dessert or cheesy bread. Call-center staffers can step in whenever the software detects customer confusion or frustration.

“It’s solving a multitude of different problems, and we didn’t realize it until after the fact,” Ahmad said of Kea’s earliest days. “Now the cashier can actually focus on the customer, the line isn’t as long out the door at peak times because they don’t have to be on the phone anymore, [and] they can just focus on the guest and providing great customer service.”

Orders are directly integrated into the point of sale, and restaurants pay 7 percent of each order for the service. So far, none of the locations that have signed up have terminated the service. To handle growing demand, Kea’s staff has swollen to nearly 100 as more restaurants join the platform. Ahmad stressed that he wouldn’t perpetually add call center employees, but focus instead on continually improving the AI system’s performance and abilities.

A calculator on Kea’s website allows restaurants to plug in average call volumes, ticket sizes and the number of phone operators at a given restaurant. Using 10 orders, a $20 average ticket and two operators, that’s an estimated monthly savings of more than $3,500 and increased revenue of nearly $4,500. Of course, your mileage will vary.

Asked about phone orders seeming retrograde as more tech filters into the restaurant world, Ahmad said improving that ordering experience could end up boosting call volume in certain restaurants by improving the experience. Also, he said restaurants have new incentives to self-originate delivery and takeout orders, rather than paying higher fees for orders coming in through third parties.

The company is focusing exclusively on franchised restaurants, especially in the QSR space, and is working directly with franchisors to explain the service’s benefits and, ideally, get their assistance in pushing this option out to their franchisees.

Out of the nearly trillion-dollar U.S. restaurant industry, Ahmad predicts $170 to $180 billion of that total still comes in through phone calls. Even as that volume likely shrinks over the coming years, he said Kea will roll out the voice AI service to other parts of the restaurant, like drive-thru order takers. It’s also in talks with third-party brands to possibly add delivery fulfillment to its offering.

Speaking more broadly about the labor challenge for restaurants in today’s economy, especially in markets like California and New York, he said his mission is softening the impact of recent changes for restaurateurs.

“That’s something that, me personally and as a company, we have this creed that you need to help save that,” he said. “We need to save the restaurant industry, in essence, and this is just the starting point.”