The Rise of the AI Drive-Thru: How Fast Food Quietly Automated Order Taking

When Kristen Charpentier pulled into the Dairy Queen drive-thru lane in Cary, North Carolina, the first clue she wasn’t speaking to a human was surprisingly straightforward: the voice coming through the speaker was simply “too nice.”

“It had no awkward communication lags or glitches—it felt totally natural,” Charpentier told WIRED. “But it was just a little too consistently pleasant to be a real person.” The voice ended every sentence with “please” and “thank you,” stayed soft and friendly through every step, and even its pitch to add extra Snickers to Charpentier’s Blizzard was perfectly polite. But as someone who’d worked Dairy Queen drive-thrus as a teenager, Charpentier knew something didn’t add up. When she pulled up to the window to pick up her order, she asked the employee point-blank: had an AI taken her order?

The worker confirmed it: Charpentier was one of thousands of customers across the U.S. who’d already interacted with an AI-powered voice at the drive-thru speaker. Data from Intouch Insight’s annual Drive-Thru Survey shows that in 2024, roughly 4% of U.S. fast-food locations used voice AI for drive-thru orders. That figure rose to 5% the following year, and early research puts the current share at 6% and climbing. For most customers, a first encounter with an AI order taker comes as a surprise, often captured on camera and posted to TikTok or Snapchat. But soon, these automated ordering systems will feel downright ordinary. For several major chains, voice AI is already the new normal.

Taco Bell announced last July that it had rolled out AI to 890 drive-thru lanes, more than 10% of its U.S. locations. Dairy Queen is rolling out the technology across 25 states, with long-term plans to bring it to every restaurant. 12% of White Castle locations already use an AI order taker named “Julia,” and all new White Castle locations open with Julia handling orders out of the gate.

“Brands that invested in and tested this heavily over the last three years have gotten the technology to a point where it’s ready for widespread use,” Sarah Beckett, vice president of sales and marketing at Intouch Insight, told WIRED. “They’re about to start scaling this massively.”

Somehow, without most fast-food customers even noticing, the era of human order takers in headsets has entered its final stretch. Welcome to the age of the AI drive-thru.

McDonald’s kicked off the trend back in 2019, when it acquired AI drive-thru startup Apprete, which the fast-food giant quickly folded into a new internal division called McD Tech Labs. When drive-thru demand boomed during the pandemic, other chains rushed to launch their own AI partnerships: Del Taco and Dairy Queen paired with Presto, Panda Express and White Castle worked with SoundHound, and Taco Bell partnered with Nvidia and Omilia. Chains called AI the future in press releases, promising it would be faster, more accurate, and better at upselling than any human could ever hope to be.

Then came the reality check. If you searched “AI drive-thru” on social media in 2023, you’d see exactly how the revolutionary technology was actually faring. One viral TikTok showed McDonald’s AI accidentally adding 260 Chicken McNuggets to an order, as the panicked customers screamed, “Stop! Stop! Stop!” The next year brought another clip showing AI adding bacon to a customer’s McDonald’s ice cream order. Other users deliberately tested the AI’s limits, including one man who ordered 18,000 cups of water at a Taco Bell.

Beyond the viral mockery, there were consistent, low-stakes failures. In a perfect world, AI order takers would be invisible—just like an extremely polite, extremely underpaid human employee. In practice, Intouch Insight found early AI drive-thrus tended to be less friendly than humans, with lower order accuracy. Automated systems still struggle to understand many accents and dialects, a problem most chains solve by keeping a human worker on standby, ready to step in when AI gets confused. The flexible back-and-forth of a fast-food order feels natural to humans; for most early AI models, it was a bafflingly complex problem.

“Restaurants ended up spending more time smoothing over the problems the AI created,” says Michael Schatzberg, cofounder of Branded Hospitality, a firm that invests in food service technology. “Everyone was saying, ‘This is terrible. Stop, put the human workers back on headsets.’” Branded Hospitality sold off its early AI drive-thru investments. Del Taco rolled back its AI test in 2024. McDonald’s eventually sold McD Tech Labs to IBM, then pulled the plug on its original AI ordering program entirely.

Not every chain abandoned their AI drive-thru efforts, though. Many—including Dairy Queen, White Castle, and Taco Bell—continued to quietly expand their tests. For the fast-food industry, AI represents far more than a new technology: it is the fulfillment of fast food’s core promise, to make getting food faster and faster while cutting costs lower and lower.

Four years into White Castle’s partnership with SoundHound AI, chief marketing officer Jamie Richardson says AI drive-thrus make employee shifts run smoother and speed up service. (Intouch Insight found that, on average, AI drive-thrus are 21 seconds faster than the industry average for human-run lanes.) He also notes that viral AI drive-thru disasters tend to be heavily overblown. White Castle’s strategy focuses on incremental improvement, rather than waiting for a perfectly infallible AI. “It wasn’t bad at the start,” Richardson says. “It’s just gotten so much better as time has gone on.”

Across the fast-food industry, momentum is swinging back toward AI drive-thru adoption. In addition to Taco Bell and Dairy Queen’s planned expansions, McDonald’s announced this summer that it is testing AI-powered ordering once again, this time partnering with Google at five U.S. locations, with a system that takes orders in both English and Spanish. Branded Hospitality has even reentered the space, recently partnering with Arc, a voice AI drive-thru company that works with chains including Southern U.S. favorite Jack’s and international chicken chain Pollo Campero.

AI drive-thru 1.0 was a flop, Schatzberg says. But now? “Companies have evolved, technology has evolved, and they’ve fixed most of those old problems,” he says.

Some customers remain unconvinced. When an automated voice started taking his order at a Wendy’s in Virginia, Schereé Reed says his dominant feeling was “bewildered.” Reed kept waiting for a human to take over; when the voice kept its distinctly robotic tone, he found himself wondering if his burger would even be served by a human, or just dropped off via conveyor belt. “I worry a lot about the ways relying more and more on AI erodes human connections, and our ability to build small moments of community,” Reed told WIRED.

Certain chains—especially those with a reputation for top-tier customer service—share these concerns. Chick-fil-A consistently tops Intouch Insight’s ratings for the friendliest drive-thru service, despite its famously long lines, and the chain has never hinted at replacing human workers with robots. Two of the fastest-growing U.S. coffee chains, Dutch Bros and 7Brew, have similarly rejected AI ordering.

“Every brand has to make choices that fit its business,” Nick Chavez, chief marketing officer at 7Brew, told WIRED in an email. “For us, hospitality starts with genuine human connection, and that's something our guests consistently tell us they value.”

Customers like Reed still prioritize that connection. “I want to be able to say hi to a human being,” he says. At White Castle, customers can request to speak with a human worker at any point during their order. But Richardson says “Julia” has already won over employees. White Castle was the only fast-food chain to crack Fortune’s top 20 list of best workplaces in retail last year; Richardson says the chain’s AI-driven approach to drive-thrus plays a role in that high employee satisfaction.

Executives claim AI isn’t stealing anyone’s job, and data so far supports that claim. Most drive-thru workers already juggle multiple tasks; while many chains would love to cut costs by replacing cashiers and cooks with robots, fully replacing human workers end-to-end is extraordinarily difficult to pull off. White Castle’s employment numbers have stayed steady throughout its AI rollout. Taco Bell recently reported that locations using voice AI actually have higher employee retention rates than those that don’t.

What chains are actually doing with AI is using it to get customers to spend more on fast food. Intouch Insight’s survey found that AI drive-thru agents use “suggestive selling”—upselling—71% of the time, compared to an industry average of 58% for human workers. Convincing customers to add a Diet Coke or extra Snickers to a Blizzard boosts profit margins without angering most guests, and AI has no shame when it comes to the upsell.

The perceived benefits for fast-food chains don’t stop there. In the near future, AI-equipped drive-thrus may be able to take orders in dozens of languages. They can recognize a customer the moment their car pulls up to the speaker, and offer suggestions based on a detailed, encyclopedic knowledge of each customer’s order history. Imagine pulling up to McDonald’s and hearing a chipper automated voice offer you a buy-one-get-one-free milkshake deal—then realizing the chain offered you the deal because its algorithm calculated you are 78% more likely to order dessert on the second Tuesday of every month.

“We never want to seem creepy,” Richardson says. For customers like Reed, that underlying unease is already hard to ignore.

Charpentier, meanwhile, sees the AI drive-thru as a net positive. And she’s in the majority: Intouch Insight reports a staggering 97% customer satisfaction rate for the technology. Charpentier says that when she worked the Dairy Queen drive-thru as a teenager, she would have jumped at the chance to have an AI coworker handle orders.

“I think it's a great thing,” Charpentier says. “There's room for human error, and I'm sure there's going to be room for AI error. But I think, as we've all seen with the spread of AI, there is more consistency with AI than there is in humans.”