The SaaS-ification of the Drive-Thru: Trading Labor Costs for Licensing Fees

The narrative surrounding the AI drive-thru has largely been one of customer friction: the viral TikToks of 260 Chicken McNuggets appearing on a McDonald's order or the chaos of a customer attempting to order 18,000 cups of water at Taco Bell. But for those tracking the money, the operational glitches are a distraction. The true structural shift is the movement of capital from human labor costs to recurring software-as-a-service (SaaS) licensing fees.

For decades, the drive-thru has been a battle of seconds and cents. The goal has always been to accelerate the transaction while minimizing the cost of the person wearing the headset. Now, as first reported by Wired, the industry is attempting to automate that role entirely. According to the annual Drive Thru Survey from Intouch Insight, the adoption rate of voice AI in U.S. fast-food lanes grew from 4 percent in 2024 to 5 percent last year, and early research indicates it has now reached 6 percent.

While these percentages seem incremental, the scale of the rollout among major players suggests a massive pivot in how these companies allocate their operational budgets. Taco Bell announced in July that it has deployed AI in 890 drive-thru lanes, representing more than 10 percent of its U.S. locations. Dairy Queen is currently implementing the technology across 25 states with the ultimate goal of a full-fleet rollout. White Castle has gone even further, with 12 percent of its locations utilizing an AI named “Julia,” and a mandate that all new locations open with the system in place.

From a markets perspective, this is a migration of the fee. The cost of a human employee is variable and subject to labor market fluctuations. The cost of an AI agent is a licensing fee paid to a technology provider. The list of these providers reads like a Who's Who of the tech world: Nvidia and Omilia (Taco Bell), SoundHound (Panda Express and White Castle), and Presto (Del Taco and Dairy Queen). Even McDonald’s, which previously sold its McD Tech Labs division to IBM and abandoned early efforts, has returned to the fold this summer, partnering with Google to test AI-powered ordering at five U.S. locations.

This shift is not without its risks. Michael Schatzberg, cofounder of Branded Hospitality, noted to Wired that early failures led some operators to spend more time fixing AI-generated errors than they would have spent taking the order manually, leading his firm to sell off early AI-drive-thru investments. Del Taco, for instance, rolled back its AI tests in 2024.

However, the proponents of the tech argue that the efficiency gains outweigh the early errors. Jamie Richardson, chief marketing officer at White Castle, told Wired that their four-year partnership with SoundHound AI has resulted in smoother employee shifts and faster service. This is backed by Intouch Insight data, which finds that AI drive-thrus beat the industry average speed by 21 seconds.

As Sarah Beckett, vice president of sales and marketing at Intouch Insight, explains to Wired, the brands that have spent the last three years testing the technology are now ready to scale massively. The objective is clear: replace the human headset with a scalable software license that can upsell with perfect politeness and never call in sick.

In the end, the customer may not notice the difference—as evidenced by Kristen Charpentier, who only realized she was speaking to a bot at a Dairy Queen in Cary, North Carolina, because the voice was “too nice’. But the balance sheets will. The drive-thru is no longer just a point of sale; it is becoming a subscription service.