Slang AI

The phone rings. Somebody finally picks up.

It is 7:48 on a Friday. The pass is backed up, two servers called out, and the host stand phone is ringing for the ninth time in an hour. At a few thousand restaurants, nobody is sprinting to grab it - because the call has already been answered, in a warm voice, by software.

That software is Slang AI. The New York company builds what it calls a "Superhost" - an agentic voice AI that picks up the restaurant phone around the clock, books and changes reservations, fields the "are you open on Memorial Day" questions, captures private-dining leads, and quietly recognizes the regular who always asks for table 12. The pitch is not that machines should run hospitality. It is that the phone, the oldest piece of restaurant technology in the building, has been quietly losing money for decades, and somebody finally decided to fix it.

"Slang makes personalization possible at every restaurant, not just a handful of elite establishments." - Alex Sambvani, Co-Founder & CEO

A missed call is a missed table.

Restaurants have a maddening math problem. The busiest moments on the floor are also the busiest moments on the phone, and a human can only be in one place. So calls go to voicemail. Reservations evaporate. The party of eight that wanted to book a birthday simply calls the place down the street that answered.

The numbers Slang cites are blunt: at some operators, up to 80% of calls went unanswered during peak hours, and roughly one in five calls arrives after the lights are off and the chairs are stacked. One restaurant group, the story goes, had 700 voicemails piled up before switching the phones over. Nobody was ignoring guests on purpose. There was simply no one free to say hello.

"Before Slang AI, 80% of calls would go unanswered because staff was too busy. Now guests get answers fast." - Charlie, Director of Operations, Crave Fishbar Group

The conventional fixes were all faintly insulting to the guest. Phone trees. "Press 1 for hours." Hold music composed, apparently, to test the limits of human patience. None of it sounded like hospitality, which is the one thing a restaurant is actually selling. The opportunity was hiding in plain sight: make the phone sound like the best host you ever met, and make that host available at 2 a.m.

Two Spotify data scientists walk into a restaurant.

Alex Sambvani and Gabe Duncan did not come from hospitality. They met as data scientists at Spotify, a company that turned listening habits into one of the most studied datasets on earth. In 2019 they left to chase a less glamorous signal: the restaurant phone line. Their bet was that voice was about to become a credible interface, and that the place it would matter first was wherever a missed call had an immediate dollar value attached.

It was, on paper, an odd career move - from soundtracking your commute to rerouting reservation calls. But the founders saw the same thing twice: messy real-world demand, and a model that gets sharper the more of it you feed it. Restaurants generate an enormous volume of nearly identical conversations. That is a nightmare for a tired host and a gift for a learning system.

"Slang has built an insurmountable advantage through proprietary data, with customer satisfaction almost unheard of in enterprise software." - Rick Lewis, US Venture Partners

One host. Every call. No smoke breaks.

Slang's Superhost is not a phone tree wearing a personality. It answers in a natural voice - the restaurant picks the accent and tone - and actually does things. It books, modifies and cancels reservations through the systems the restaurant already uses. It routes a private-dining request to the events manager instead of dropping it. It remembers the guest who has called eleven times this year. Setup, the company says, usually takes under 30 minutes, which is less time than it takes to train a new host to find the reservation book.

"They had 700 voicemails before launching Slang AI. Now they don't worry about them." - Dustin, Weimann Maclise Restaurants

It also plugs into the plumbing operators already trust - OpenTable, SevenRooms, Yelp for reservations, and Tripleseat for events - so the AI host is not a sealed island. It is a colleague that happens to never sleep, and never gets that glazed look at the end of a double.

The receipts, plotted.

Hospitality is a notoriously hard sell for software - margins are thin and patience for gadgets is thinner. Slang's argument is that the math is unusually clean: a call answered is revenue kept. Operators report up to 20x ROI and a roughly 2x lift in phone reservations, which is the kind of number that survives a skeptical GM's eye-roll.

The customer list runs from steakhouse chains to neighborhood groups: Texas de Brazil, Carmine's, Riot Hospitality Group, DineAmic Hospitality, Founding Farmers, Crave Fishbar and Merchants Hospitality among them.

Hospitality that scales without going cold.

The easy version of this story is "AI replaces the host." Slang tells a different one. The point of answering the phone, the founders argue, is to free the humans on the floor to be human - present with the guest in front of them instead of lunging for a ringing handset. Great service has always been a scarce resource, available mostly at the restaurants that could afford to staff it. Slang's mission is to make that level of attention ordinary rather than rare.

"The phone never died. It was just understaffed." - The Slang AI thesis, paraphrased

There is an irony worth savoring: the most futuristic thing Slang built is a better way to do the single most analog task in the restaurant. No app to download, no QR code to squint at. Just a voice that picks up and gets it right.

The data flywheel starts spinning.

With the Series B, the bet gets bigger. Every call teaches the system something - about menus, about peak hours, about what guests actually ask for - and that proprietary record of ~25 million conversations is the moat investors keep pointing at. The plan is to push deeper than reservations: operational insight, guest intelligence, and eventually the same playbook for hotels and venues, not just dining rooms.

Whether voice AI becomes the default front door for hospitality is still an open question, and skeptics are right to ask whether a warm-sounding bot can hold the line as expectations rise. But the direction of travel is hard to argue with. The phone is not going away. The question is only who answers it.

"Personalization at every restaurant - not just the famous ones." - Slang AI, on the next chapter

So picture that Friday again. 7:48 p.m. The pass is still backed up and two servers still called out. The phone still rings. But this time nobody flinches, because the call is already handled - the table booked, the regular greeted by name, the birthday party saved from defecting down the block. The rush did not get quieter. The restaurant just stopped dropping the calls it used to lose. That is the whole product, and it is plenty.