Bilingual Customer Service in Liquor Stores: Why Spanish Matters and What Most Stores Get Wrong

A Spanish-speaking customer walks into your liquor store, points to the wall of reds, and asks for "un vino tinto suave para la cena." If the clerk speaks English only, three things happen — and none of them are good. The customer takes a guess, the clerk gestures vaguely, or the customer walks out and tries the store across the street. I have watched this play out in person more times than I can count.

This is not a diversity-and-inclusion essay. It is an operator problem, and it is one of the cleanest cases I know for an AI assistant in an indie store. Here is how to think about it.

The business case before the moral case

Hispanic shoppers in the United States spend a real and growing share of off-premise alcohol dollars. I am not going to throw a fabricated stat at you — go pull your own POS data and break out high-Spanish-surname zip codes if you want the local picture. What I will say is that in the indie stores I work with in Southern California, Texas, and Arizona, the Spanish-speaking shopper is somewhere between 20 and 60 percent of weekday foot traffic, depending on the neighborhood.

Now look at staffing. Most independent liquor stores I have walked into have one clerk on shift. That clerk's first language is whatever it happens to be. If the customer's first language does not match, the friction is real and it gets paid for in lost basket size, not lost transactions. Customers do not walk out empty-handed. They walk out with the cheapest sure thing, which is rarely what they wanted.

That gap — between what the shopper wanted and what they actually bought — is what bilingual service closes. It is not a feel-good metric. It is a basket-size metric.

The "vino tinto suave" example

Let me unpack the example from the top. "Vino tinto suave" translates literally as "soft red wine." In Mexico and most of Central America, "suave" in the wine context usually means low tannin, low acidity, fruit-forward. Think malbec, merlot, grenache, garnacha, some pinot noirs. It does not mean sweet, though plenty of customers will accept sweet if you read it that way.

Now imagine you are a clerk who took two semesters of high school Spanish. You hear "suave," your brain jumps to "smooth," and you point at the smoothest-looking thing on the wall — which might be a bone-dry cabernet that the customer is going to hate. The customer takes it home, does not enjoy it, and either returns it or quietly moves their alcohol spend to somewhere else.

A native Spanish speaker would have grabbed a malbec under fifteen dollars and explained it in one sentence. That is the gap. Multiply it by every "vino blanco para mariscos" or "tequila para sipping" or "cerveza ligera que no sea Bud Light" question that walks through your door.

What bilingual signage misses

The standard answer to this problem in indie retail is bilingual signage. Spanish category headers, maybe a Spanish-translated promo sheet at the register. Better than nothing. But it solves the wrong problem.

Bilingual signage helps a customer who already knows what they want and just needs to find it. It does nothing for the customer who has a question, a budget, a use case, and no language to express it. The "vino tinto suave" customer is not going to read your sign. They are going to ask a human, and if no human in the building speaks their language, they are going to compromise.

Signage is static. Conversation is not.

What bilingual staff costs

Let me give you the math on the obvious solution — hiring a bilingual clerk — because most operators have already done it in their head and decided it does not work.

A bilingual clerk is worth maybe two to four dollars more an hour than a monolingual one. If you pay that across forty hours a week, fifty weeks a year, you are looking at four to eight thousand dollars in incremental wages per bilingual clerk. That is for one shift. If you want bilingual coverage across both shifts, double it. You also have to actually find that person — and good luck in a tight retail labor market.

If you are running a single-store operation with two or three full-time clerks, an AI assistant that handles bilingual product questions across every shift is meaningfully cheaper than one bilingual hire. That is before you count the fact that the AI never calls in sick. It is a real comparison and the math does not require any bullshit assumptions.

What a kiosk actually does in this scenario

A bilingual kiosk like Remi does the boring, repeatable part of bilingual service well. The customer walks up. The greeting is in English by default but the kiosk listens for Spanish and switches without a button press. The customer says "vino tinto suave para una cena con carne," and the kiosk responds in Spanish with two or three actual bottles the store actually has, with prices, and routes the customer to the aisle.

The clerk is freed up to do the things only a human can do — ID checks, register, complicated returns, conversations with regulars. That division of labor is the whole point.

A few things to be specific about, because vendors get this wrong:

The kiosk has to know the store's actual stock. Recommending bottles you do not carry is worse than recommending nothing. Inventory feed integration matters. See POS data privacy and what kiosks should read.

Spanish has to be native, not translated. A model that speaks Castilian Spanish to a customer from Jalisco sounds wrong. Mexican Spanish is the default for U.S. indie liquor; the assistant should default to neutral Latin American Spanish unless told otherwise.

Voice matters more than text. A lot of the customers most underserved by English-only service are also customers who would rather speak than type, especially older adults. A kiosk that requires typing is solving the wrong half of the problem.

Hispanic-owned vs. Hispanic-serving stores

Worth distinguishing two different operator situations.

Hispanic-owned stores often already have a bilingual clerk — frequently the owner. The kiosk's value here is more about coverage when the owner is not in the building, plus consistent product knowledge for new hires who may not match the owner's depth.

Hispanic-serving stores — owned by someone whose first language is not Spanish, in a heavily Spanish-speaking neighborhood — have the bigger gap. This is where the friction is most visible and the kiosk pays for itself fastest.

Both groups benefit. The first group does not always realize how much.

The right way to roll this out

If you are an operator thinking about this, here is the order I would do it in.

1. Pull a week of receipts. Look at average basket size by hour. If you have any way to overlay neighborhood demographics or surname data, do it. Find the gap between your weekday afternoon (typically the most diverse) and your peak weekend evenings.
2. Talk to your clerks. Ask them how often a customer asks a question they cannot answer because of a language gap. The number is bigger than they will initially admit.
3. Pilot a single kiosk with bilingual support for sixty days. Read the conversation logs. Look for Spanish-language interactions and whether they converted.
4. Decide based on the actual data, not the vendor's deck.

If you want to talk through what a pilot looks like, that is what we do.

Frequently asked

Will my Spanish-speaking customers actually use a kiosk?

Yes, more than English-speaking customers in our pilots. The voice-first experience is more familiar than typing for many older shoppers, and the privacy of asking a screen rather than a stranger lowers the embarrassment threshold for "I do not know what this is" questions.

What if the customer mixes languages?

Spanglish is normal. The model handles it. The interesting case is a Spanish question with English brand names ("¿tienen Maker's Mark?") — the kiosk should answer in Spanish without translating the brand. Test this when you demo any vendor.

Does this work for languages other than Spanish?

Technically yes — the underlying model is multilingual. We default to English and Spanish for liquor stores because that is the dominant gap. Mandarin, Vietnamese, Tagalog and Korean are all technically supported but we have not productized those yet.

Can the kiosk recommend wines from specific regions my Spanish-speaking customers prefer?

Yes — and this is the part that surprises operators. Mexican wine (Valle de Guadalupe), Argentine malbec, Spanish rioja, Chilean carmenere — these are categories your Spanish-speaking customers may know better than your clerk does. The kiosk can match preferences to your actual stock without making the customer feel talked down to.

What does this cost compared to hiring a bilingual clerk?

Cheaper, when you actually do the math. See the wages section above. The clearer comparison is at pricing.