Thanks to Nvidia, AI will soon take your order at Taco Bell and Pizza Hut

Yum Brands, owner of Taco Bell, Pizza Hut, KFC and Habit Burger & Grill, announced today it is partnering with Nvidia to build a range of new AI-driven services in its restaurants.

The first—AI-powered voice-ordering at the drive-through lane and on the phone—was built by Yum’s developers using tools and development frameworks from Nvidia, and will begin rolling out at 500 restaurants across the portfolio in the second quarter of this year.

The ultimate goal is to move all orders through digital channels instead of human order-takers, an effort Yum says will boost sales. Yum will also use AI to enhance a number of the company’s internal operations.

Other planned changes include the use of computer vision to spot fumbled orders and AI that filters internet chatter on the restaurants for useful feedback for their managers, all of which Nvidia software will be deeply involved in helping build.

Terms of the deal weren’t disclosed.

The partnership comes at a pivotal time. Yum and other quick-service-restaurant chains, like McDonald’s, have been leaning into more digital experiments for efficiency gains and improved customer satisfaction as inflation squeezes low-income diners. In some cases, the results have been mixed.

Joe Park, chief digital and technology officer at Yum Brands.

Chief Digital and Technology Officer Joe Park said Yum’s decision to use Nvidia tools to build in-house, rather than rely on third-party vendors with prepackaged apps, was a pivot in strategy for the fast-food giant, which operates over 61,000 restaurants.

“We want to own the intellectual property. We want to own the technology," Park said. “That’s a shift in our strategy as we think about AI."

Its recently announced “Byte by Yum" proprietary platform will house the new apps, said Park, who is also president of the platform. Third-party vendors, however, will always have a role in certain areas, the company said.

Andrew Sun, Nvidia’s director of AI business development for retail, said the partnership was unique for Nvidia partly because of Yum’s advanced in-house tech talent. The Yum team was primed to take advantage of Nvidia tools and application frameworks like its NIM microservices, which is designed to aid AI development.

Yum’s 2,000 technologists will use Nvidia software to build apps that leverage open-source models and take full advantage of the computing power of Nvidia’s GPUs. Nvidia will also provide guidance and advice.

Building in-house with Nvidia helped bring the cost down, Park said, since the AI chip giant helped Yum maximize every bit of compute on Nvidia GPUs. As a result, the apps scale better and seamlessly fit into Yum’s tech stack, he added.

Voice-ordering has been a priority for the chain for some time as it works to receive 100% of its orders through digital channels rather than through humans. Currently it is above 50% including orders that come through its app or online, up from 19% in 2019.

Consumers end up spending more when they buy via digital channels, Park maintains, because the restaurant can upsell, personalize and entice eaters through notifications. And that is critical as the fast-food industry faces a pullback amid inflation concerns and churning economic policy.

“We love everything about digital sales," Park said. “The more we can create channels, whether it’s voice-ordering, whether it’s going to be tied to our kiosks, whether it’s going to be tied to investments in AI to help encourage customers to go to our mobile app. I think all of these things add up to helping in these economic environments."

But in-restaurant workers won’t disappear, Park said. Instead, they will focus more on customer service, for instance, helping people with orders. “It’s a different form of service and hospitality than I think we’ve seen in the past," he said.

Yum is also looking to enhance its proprietary mobile app, used by restaurant managers at more than 20,000 restaurants to track and manage operations. Yum is building a feature that ingests online customer reviews and turns them into proactive suggestions to restaurant managers.

Improving order accuracy is also an area where Nvidia will lend value, Park said. Yum is currently evaluating whether existing CCTV cameras can provide images sharp enough for computer vision to determine whether the food received is what was ordered or whether it’s missing any ingredients.

“Order accuracy is a big problem that a lot of [quick-service restaurant] companies face," Park said.“We think there’s a lot of runway here to innovate here," added Nvidia’s Sun.

Write to Isabelle Bousquette at isabelle.bousquette@wsj.com