A viral video from 2023 showed a frustrated Wendy’s customer at a drive-thru struggling to place an order through an automated AI system. The machine instructed her to move to the next window without allowing her to add items to her Dave’s Double burger order. An employee eventually intervened to correct the mistake, but the incident went viral. A sign on the drive-thru speaker warned customers to avoid filler words like “um” and “uh” because the AI was still learning and would otherwise interrupt users. The video’s caption—“What a joke. I CANNOT, AI in the drive-through. See you never, Wendy’s”—became one of many social media posts highlighting customer and employee frustration with AI in restaurant operations.
The technology has been deployed for tasks ranging from staff forecasting to drive-thru ordering, but its implementation has faced persistent criticism. Lu Lu, a Temple University professor specializing in food service consumer behavior, notes that AI adoption in hospitality, both in customer-facing roles and behind the scenes, has drawn significant scrutiny. The core issues arise from technology that fails in critical areas, including inventory management, order fulfillment, and even generative menu imagery. Visible failures do more than create operational problems; they erode brand loyalty.
Public-facing mistakes, such as AI-generated menu images that appear unappetizing, have triggered consumer backlash. Nik Allen, global insight manager for consumer foodservice at Euromonitor International, warns that restaurants face heightened risks when AI failures become visible to customers. Logistical errors, such as incorrect inventory counts or mislabeled items, create operational disruptions, but the largest threat lies in public-facing mistakes. “Public-facing agentic failures are the biggest threat to brand loyalty,” Allen said, “where these can be perceived less as a futurist experience and more as a cost-cutting measure.”
Major chains pull back after early missteps
Starbucks discontinued its AI inventory management system after just nine months, citing occasional miscounts or mislabeled items. McDonald’s ended its automated order-taking partnership with IBM in 2024, though it later announced plans to test Google-powered drive-thru technology at five of its U.S. locations. The company is also advancing with ArchIQ, an AI-powered restaurant operating system designed to handle drive-thru voice commands, automated inventory tracking, and equipment monitoring.
Despite these setbacks, the broader trend shows no signs of slowing. The challenges extend beyond technical failures, employees report frustration over increased labor demands and anxiety about job displacement.
A Pew Research Center poll from June found that 52% of Americans are more concerned than excited about AI’s growing role in daily life, with job loss fears driving much of the unease. When restaurants fail to explain why they’re adopting AI, trust erodes further, leading to lower morale among staff.
Even as concerns grow, investment in AI isn’t declining. The same Deloitte survey revealed that 63% of executives use AI daily to enhance customer experience, while 26% are testing limited implementations. Nearly 80% plan to increase AI spending in the next fiscal year, betting it will improve operations, loyalty programs, and overall efficiency.
The most successful AI applications so far have been in demand forecasting and inventory management, areas where large data sets and human error create inefficiencies. AI has also been effective when used to set worker scheduling, though its effectiveness declines sharply when it interacts directly with customers.
Experts recommend a cautious approach: begin with small-scale deployments, focus on back-of-house functions, and involve employees early to address their concerns. The objective, as Lu Lu puts it, is to use AI to “reduce those frictions where humans struggle and smooth out the process”, not introduce new problems.
AI’s limited success in back-of-house roles
Yet even in these areas, challenges persist. The Deloitte survey from late 2024 revealed that 48% of restaurant executives remain concerned about managing AI-related risks, with 20% citing intellectual property issues and 16% worried about data misuse. Evert Gruyaert, Deloitte’s U.S. restaurants and food service leader, emphasizes that the primary hurdle isn’t a reluctance to invest but rather identifying the right use cases and scaling them responsibly. “That suggests the challenge may be less about willingness to invest and more about identifying the right use cases and scaling them responsibly,” Gruyaert said.
Lessons from early adopters
Starbucks’ decision to abandon its AI inventory system after nine months highlights the consequences of premature scaling. The system occasionally miscounted or mislabeled items, leading to operational inefficiencies and customer frustration. Meanwhile, McDonald’s shift from IBM’s automated order-taking to Google’s drive-thru tech, and its plans to deploy ArchIQ, demonstrates the industry’s willingness to experiment, even after setbacks. Allen describes the current phase as “the frontier phase of agentic applicability,” suggesting that AI’s potential remains unproven in many areas.
Experts agree that restaurants should proceed with caution. Pilot programs allow chains to test AI tools in controlled environments before full rollout, helping to identify flaws early. Lu recommends starting with back-of-house functions where AI can reduce human friction, such as scheduling or inventory, before expanding to customer-facing roles. “In theory, at the early stages of adoption, AI should be used to really reduce those frictions where humans struggle and smooth out the process,” Lu said.
