Phone lines are incredibly vital for restaurants.
Data shows that 71% of calls are directly tied to revenue. A working phone line is a must-have communication channel.
When guests call, these moments are opportunities to leave a lasting impression. A top-notch restaurant host achieves this by adapting to each guest's preferences and needs.
Want to embody hospitality to guests? Then you’ll need to speak their language.
For many restaurants in the U.S., a significant number of guests prefer Spanish. The team at Slang AI saw an opportunity to develop bilingual support for both English and Spanish speakers. Let’s explore the process of this latest enhancement and why this is a significant release.
The Gap in Guest Experience for Spanish Speakers Let’s take a quick look at the numbers. In the U.S. alone:
Spanish is the second most spoken language, after English. 13.7% of people speak Spanish at home.Over 35 million adults speak Spanish fluently. The U.S. is incredibly diverse, with millions not speaking English most of the time in their lives. Whether it’s because of the community they’re a part of or a personal preference, countless restaurants serve guests who speak other languages.
We know this because our data tells us so.
Rebecca is an expert in conversation design for voice AI experiences and had a primary role in developing and launching bilingual support along with the rest of the team at Slang AI. We interviewed Rebecca to learn more about what it took to make a truly bilingual agent a reality.
“We’ve been observing millions of restaurant calls for six plus years,” said Rebecca Evanhoe. “People do want to speak to our voice concierge in Spanish, and we've seen a lot of calls where people attempt to speak in Spanish.”
In cities like Miami, Chicago, and Los Angeles, where Spanish is a dominant language, supporting bilingual guests could mean capturing up to a third of all incoming calls.
Despite the large Spanish-speaking population in the U.S., restaurant phone systems haven't kept up. Guests are often met with frustrating prompts, asked to press a number, or told to wait for an English-speaking staff member.
“Unfortunately, traditional call systems are pretty poor experiences for Spanish speakers,” Rebecca explained. “Our testers had low expectations based on past experience. That's why we include a small nudge telling them they can speak in Spanish, so they know it will work for them.”
Building Bilingual AI to Embody True Hospitality Creating a bilingual conversational AI experience requires balancing language nuances to sound natural to guests. But it’s not as simple as switching languages and using direct translations.
The team knew that literal translation wasn't enough. They spent months researching dialects, conducting usability tests, and refining the technology to understand and adapt the nuances of Spanish as it's spoken across the U.S.
“One crucial piece of feedback we got was that it was a little too literal a translation,” Rebecca shared. “We didn’t want to lose the qualities to preserve a hospitable tone. That led to intentionally training our voice concierge to be extra conversational and warm.”
That feedback hit on something crucial: hospitality must come first.
“It all comes down to hospitality,” Rebecca said. “If somebody is most fluent in Spanish, then speaking to them in their language is respectful and helpful.”
Beyond tone, the team had to solve numerous other practical details. This level of detail-obsessed refinement ensured that the Spanish experience wasn't just functional, but seamless and trustworthy.
For example, should an address like “55 Main Street” be translated? (Spoiler: the street name shouldn’t.) In this case, over-translating risked confusing guests who were looking for an actual sign.
When Human Expertise Meets AI While AI continues to advance rapidly, Slang AI's team knew better than to rely on technology alone. Translation models by themselves couldn't deliver the warmth and nuance that restaurants actually need.
Consider a simple scenario: when a caller asks “¿Tienen mesas disponibles?” (Do you have tables available?), a direct LLM translation might respond with “Sí, tenemos tablas disponibles.” This is technically correct, but using “tablas” (boards/charts) instead of “mesas” (dining tables). While understandable, it sounds robotic and unnatural.
“You need people who are fluent, who have lived experience of that language, to tell you their expectations,” Rebecca emphasized. “If you don't speak the language, you shouldn't assume you know where the quality bar is.”
That's why, in addition to user testing with actual native speakers, the team enlisted the help of external conversation design experts with fluency in Spanish to assist with linguistic reviews. Human oversight was essential in shaping an AI that feels as personal and warm as a staff member answering the phone.
The Future of Multilingual AI in Restaurants Slang AI's launch of Spanish bilingual support is a technical milestone as well as a statement about where hospitality is heading. As Rebecca Evanhoe suggests, the bar will only get higher from here.
Some day, guests will expect the Spanish experience, or any multilingual interaction, to be as polished, natural, and human as the English version.
For restaurant operators, embracing voice AI means considering the unique needs of their guests and utilizing solutions that can meet those expectations. The right solution will capture more revenue, strengthen guest loyalty, and embody true hospitality.
Offering bilingual support shows what's possible when technology and cultural understanding come together. To speak the language of hospitality, you have to communicate in their actual language.
