Taco Bell Has Processed 2M+ Drive-Thru Orders Using AI; Now In Use at 300+ US Stores

November 7, 2024 by Dave Haynes

Hat tip to kiosk expert Craig Keefner for pointing to this …

Anyone wondering if voice ordering in QSR drive-thrus would mature and would get adopted by consumers can reference what Taco Bell is doing in the U.S. The CFO of Yum! Brands, which owns Taco Bell, KFC and Pizza Hut, said on a Q3 earnings call that more than 2 million orders have been processed using voice AI and the tech is actively being used at more than 300 stores.

“Drive-thru voice AI continues to scale across our network, with many franchisees eager to test this new innovation,” Chris Turner said on the call, reports PYMNTS.

“To date, we have processed over 2 million successful orders with the system now in place in over 300 Taco Bell U.S. stores, making Taco Bell the largest QSR voice AI brand in the world.”

“For loyalty, Taco Bell is using its connected ecosystem to allow loyalty consumers to identify themselves at the drive-thru and kiosk, enabling personalization of their ordering experience and earning and redeeming of loyalty rewards. This was rolled out to 160 stores in Q3, and we’re encouraged with early results, which clearly show an increase in sign-ups and in daily loyalty transactions, all without an impact to speed of service.”

Turner also detail on the call how Yum! is actively focused on acquiring, building and scaling proprietary platforms “to gain control over their data, streamline innovation, and leverage cost efficiencies through scale. The second phase aims to maximize the value of these platforms by using artificial intelligence (AI) and the company’s extensive data assets.”

In other words, they’re bringing systems in-house. I was on a call earlier this year that had some QSR IT people asking about digital signage and menu display applications, and it was unstated but pretty obvious the company was not intending to go with a third party but wanting to DIY something.  Sometimes they find out doing it well is hardware than expected, but that was an assertion more easily made five years ago than now – with all the APIs, cloud services, developer tools and especially AI.