Hostie raises $12M Series A2 million guest conversations400,000+ covers bookedPlans from $199 per locationBuilt inside a San Francisco restaurantHostie raises $12M Series A2 million guest conversations400,000+ covers bookedPlans from $199 per locationBuilt inside a San Francisco restaurant

Company profile / AI + hospitality

The Phone Wouldn't Stop Ringing. Hostie Turned That Restaurant Headache Into a $16 Million AI Company

A restaurateur and an AI engineer started with one painfully ordinary problem: the phone ringing during dinner service. Their answer now handles millions of guest conversations - but its real test is whether automation can make a restaurant feel more human, not less.

The first product failure in Hostie's story was not artificial intelligence. It was a cordless phone. During dinner at Back to Back, Randall Hom's Italian-Californian restaurant in San Francisco, that phone kept making an impossible demand: serve the guest in front of you, or answer the guest who might book tomorrow. Calls, emails, reservations, takeout requests and private-event questions piled into a noisy little tax on hospitality.

Hom had the unusually useful combination of being annoyed and qualified. Before becoming a restaurateur, he had worked in product design at Instacart. He called his longtime friend Brendan Wood, an AI engineer then working at Lindy.ai. Together they made a narrow bet: instead of asking restaurants to become technology companies, give the ringing phone a competent colleague.

That colleague became Hostie, founded in 2024. Its conversational agent answers calls, texts and emails, handles predictable questions, books or changes reservations, takes orders, captures event leads and produces transcripts. Staff can watch, receive alerts and take over. In July 2026, Hostie announced a $12 million Series A led by Obvious Ventures, bringing its total funding to $16 million.

Hostie co-founder and CEO Randall Hom inside a restaurant
Randall Hom, seen in his natural habitat: close enough to a restaurant phone to remember why he built Hostie.

A host for the part nobody sees

The product lives in the unglamorous minutes around a meal. A diner calls to ask about corkage. Another is running late. Someone wants a table for six next Friday; another wants to move a 7:30 reservation to 8. A parent needs an allergy answer. A company is shopping for a 40-person dinner. Each exchange is simple until 12 of them arrive during service.

Hostie can answer in more than 20 languages and use a restaurant's configured tone and rules. When a booking is involved, it can check real-time inventory through systems such as OpenTable, Resy, Tock or SevenRooms. For takeout, integrations with Toast, Square and Olo push an order into the systems already running the restaurant. Tripleseat and Perfect Venue connect private-event leads to the people who can close them.

This integration strategy is the difference between a talking FAQ and operational software. A generic bot can tell someone the restaurant closes at 10. A useful restaurant concierge can notice there is no table at 7, offer 7:45, book it, record the guest's note and send confirmation. Hostie does not try to replace the restaurant's stack; it translates conversation into actions inside that stack.

“We built Hostie to help teams keep up with that demand while staying focused on the people in front of them.”Randall Hom, co-founder and CEO

What it costs - and what pays for it

Hostie sells subscriptions per location. Its Essential plan starts at $199 a month for walk-in restaurants that want a unified place for calls, texts and emails, live transcripts and basic customization. Premium begins at $399 and adds full reservation integrations, takeout ordering, deeper prompts, 20 languages and around-the-clock support. Hospitality Plus starts at $599, while enterprise groups negotiate custom terms.

The smallest public plan
$199 / month

Per location. The simple break-even question: can answering the calls you currently miss recover enough bookings, orders or event leads to cover it?

The buyer is not really purchasing a synthetic voice. The buyer is purchasing recovered attention and fewer abandoned intentions. For a busy full-service restaurant, one private dinner or a handful of covers can clear the monthly price. For a quiet cafe that receives three calls a day, the arithmetic may be much less charming.

That distinction matters because voice AI is crowded. Slang.ai, SoundHound, PolyAI, Newo and RestoHost all chase parts of the restaurant-call market. A conventional answering service, voicemail or another employee can solve it too. Hostie's claim to a defensible position is vertical depth: restaurant operators helped shape it, its vocabulary is narrow, and its integrations cover the places where a conversation becomes revenue.

The phone numbers behind the phone

Hostie says it grew from five pilot restaurants in 2024 to more than 200 locations by November 2025, then to hundreds of partners nationwide. Ahead of its Series A, the company reported 10-fold revenue growth in a year, more than 2 million guest conversations, 24 million messages, over 400,000 covers booked and more than 50,000 private-event inquiries supported.

2M+Guest conversations managed
400K+Covers booked
50K+Private-event inquiries

These operating figures and the customer results below are published by Hostie and its customers. They are useful directional evidence, not a substitute for a restaurant measuring its own baseline and results.

The customer stories are more tactile. At The Stinking Rose Group, Hostie reported handling 24,000 calls in six months, making 4,700 reservations, resolving 80 percent without a transfer and returning 403 hours to hosts. The Slanted Door's Napa location reported over-the-phone covers rising 56 percent and walk-ins rising 61 percent. Belotti Ristorante said the system resolved 90 percent of calls without staff and coincided with a 15 percent increase in reserved covers.

A lively restaurant dining room, the environment that inspired Hostie
The dining room has better things to listen to than a ringing phone. Hostie's whole thesis fits in that sentence.

At DACHA Projects, the more interesting problem was linguistic. A large share of guests spoke Russian, while owner Suki Skye was receiving calls at 6 a.m. and midnight. After adopting Hostie, DACHA reported the system handling 76 percent of calls, with reservations up 23 percent and walk-ins up 32 percent. The multilingual feature was not decoration; it matched the actual room.

What changed their minds

Restaurant people are paid to distrust anything that could make a guest feel processed. The shift came from trials in recognizable restaurants and from a product that did not demand a new customer behavior. Older guests did not need to download an app. They could make the same phone call they had always made. The Stinking Rose's Dante Serafini described the benefit plainly: someone answers all the time, the same way.

For Hom, the proof arrived first in his own restaurant. For Flour + Water, Bodega, Slanted Door and other early partners, it came through calls resolved without dragging a manager off the floor. The technology earned permission by taking the repetitive part, leaving exceptions visible, and giving staff an escape hatch. That is a more persuasive adoption story than telling a restaurant to “embrace AI.”

The playbook worth stealing

Founders can copy Hostie's sequence without copying its product. Begin with a workflow you can observe at close range. Use your own operation as the first test bench. Recruit a small set of respected design partners - Hostie began with five - and measure outcomes in the customer's language. Restaurants care about covers, calls, hours and event leads, not model benchmarks.

Then integrate before expanding. Hostie's wedge was answering the phone; its value grew when the answer could become a reservation, order or qualified lead. The company widened from voice into text and email only around the same job: keeping guest communication from falling through the cracks. That is disciplined adjacency, not a random feature buffet.

There is a distribution lesson too. Customers such as Flour + Water, Slanted Door and State Bird Provisions carry more credibility with operators than generic software testimonials. Hostie's investors now include restaurant operators alongside funds and former executives from OpenTable and Yelp. In vertical software, domain trust can be both product research and sales channel.

Where the recipe breaks

Voice automation works when the question repeats, the answer is current and the next action is defined. It becomes fragile when the menu data is stale, reservation rules are poorly configured, calls are rare, or the conversation involves anger, grief, accessibility or unusual judgment. A perfect transcript of the wrong answer is still the wrong answer.

Good conditionsHigh call volume, repeated questions, maintained restaurant data, connected booking systems and a clear human handoff.
Poor conditionsLow call volume, constantly improvised policies, sensitive conversations, weak integrations or no one assigned to escalations.

Restaurants also need to decide how transparent the voice should be and monitor what it says. Hospitality contains nuance that does not fit neatly into a prompt: the regular who expects a familiar greeting, the anxious caller planning a proposal, the allergy question that deserves conservative escalation. Hostie's best role is not to impersonate the best host in the building. It is to keep that host available for moments that require one.

A narrow wedge into a large room

Hostie sits between communications software, restaurant operations and vertical AI. The initial market sounds almost quaint - people still call restaurants - but the underlying job is broader. Guests switch among calls, texts, Google Maps, Instagram, email and reservation platforms. Restaurants inherit the burden of making all those doors feel like one entrance.

The Series A will fund product development and leadership growth. Hostie's stated direction is a wider guest-engagement layer across every channel. That expansion brings opportunity and danger. The more of the guest relationship the company handles, the more valuable its context becomes; it also assumes more responsibility for data quality, privacy, tone and failure recovery.

The cleanest version of the Hostie thesis remains the original one. At 7 p.m., the phone rings while a host is greeting a table. Software takes the routine call. The human looks up. For a company built around artificial conversation, its best outcome is a real conversation happening somewhere else.

Keep ringing

Explore the product, the people and the live demonstrations through Hostie's public channels.