How Tire-Shop Chains and Pizza Franchises Quietly Became America’s Bilingual AI Pioneers

Around four in the afternoon, you’ll hear something a little odd before you see it if you walk into a tire shop on Houston’s eastern edge. When a customer’s teenage son grabs the receiver, a voice on speakerphone greets them in leisurely Spanish, inquires about the truck’s make and model, and then abruptly switches to English. The counter clerk remains unflinching. Nobody else does either. The voice is not human. For nearly a year, the bilingual AI has been responding to that query.

There isn’t much written about this aspect of the AI narrative. The real testing ground for bilingual conversational AI has been quietly operating in strip-mall franchises throughout Texas, Florida, Southern California, and the suburbs of Chicago, while the media chases frontier models and trillion-dollar valuations. chains of tires. Pizza establishments. drive-thrus for oil changes. These companies are not typically associated with technology, let alone linguistic engineering.

Even though the founders in San Francisco hardly ever admit it, there’s a reason it started here. For many years, franchise owners in places like Houston, Phoenix, and Miami have had to deal with bilingual clientele. At lunchtime, the phone rings in Spanish; at dinnertime, it rings in English; occasionally, the opposite is true. Missed calls result in missed sales. Missed calls also became intolerable during the post-pandemic staffing shortage. These owners may have been drawn to the technology more out of necessity than ambition.

The early systems were difficult to use. When Domino’s tried voice ordering years ago, the results were inconsistent, which made people wonder if they would put up with a machine on the other end. However, around 2023, something changed. The models became more adept at code-switching, the messy, spontaneous practice of blending Spanish and English that characterizes a large portion of American discourse. Last year, a San Antonio regional pizza franchisee told a trade magazine that his bot can now handle “el medium pepperoni con extra cheese” without hesitation, and that within weeks, customers stopped leaving comments.

As this develops, it’s difficult to ignore how different the rollout appears from anything Silicon Valley usually produces. Launch events don’t exist. No flashy demonstrations. Customers at a Phoenix Discount Tire store either notice or don’t when a new system is implemented. The economics are straightforward. The bot earns its keep if it schedules an appointment. The franchisee cancels the agreement by Friday if it doesn’t.

Investors appear to think that this market will quietly grow to be very large. In the past 18 months, a number of the startups that support these chains have raised significant sums of money, frequently with little media coverage. In their pitch decks, they discuss “verticalized voice AI,” which is a sterile way of saying that they have figured out how to take orders from a fatigued man who switches between two languages mindlessly at 9:47 p.m.

Walking through these neighborhoods gives the impression that something culturally significant has happened almost by coincidence. A Spanish-speaking grandmother in Cloverleaf, Texas, most likely spoke to an AI for the first time when she ordered rotors for her son’s truck. Years ago, Tesla encountered similar concerns about consumer adoption, and the response was the same. When technology resolves a genuine issue, people embrace it. When it doesn’t, they disregard it.

Whether the big chains will continue to develop this internally or return it to the platform behemoths once those platforms catch up is still up in the air. But for the time being, the person taking your large pepperoni order on a Tuesday night may be the most proficient bilingual AI in America.