Zest adds an AI restaurant concierge that books tables by text

Zest is moving its restaurant discovery experience into text messages. According to TechCrunch, the startup has introduced an AI-powered restaurant concierge that can suggest places to eat, handle group preferences, run a vote, and book a table from the same conversation.

The launch matters because it shows how consumer AI agents are starting to replace the app-opening ritual for everyday tasks. Instead of searching, filtering, comparing availability, and switching to a booking service, Zest is trying to make dinner planning work like a text thread.

A dining agent built for the group chat

Zest’s new concierge is designed to be contacted by text at (212) 470-6067, according to TechCrunch. Users can ask for restaurants based on cuisine, craving, occasion, dietary needs, or neighborhood, and can follow up with practical questions about a venue.

The agent can also be added to a group chat. In that setting, Zest says it can combine the preferences of several people, suggest restaurants that fit the group, and use an in-thread voting feature to resolve indecision. Once the group settles on a choice, the concierge can move on to booking.

That group-chat angle is the most useful part for busy freelancers, small teams, and marketers planning client dinners or informal meetups. The pain point is rarely just finding a restaurant; it is getting five people to agree on a place, time, cuisine, location, and reservation slot without a long back-and-forth.

Zest’s agent can answer questions such as whether a restaurant is likely to be busy, whether it offers Wi-Fi, and can send a menu to help users decide, according to the report.

Why Zest thinks text beats the app

Zest launched earlier this year as a restaurant discovery app, but the company is already treating text as a primary interface. Founder and CEO Mario Gomez-Hall told TechCrunch, “The agent is the product now,” describing a shift from app navigation to conversational action.

The service builds on Zest’s existing personalization system. The app connects to users’ credit cards and uses dining history to understand taste. That data is imported through Plaid, which is used by banks and fintech apps. Zest says it focuses on transactions in food and drink categories when building its personalized maps and lists.

That gives the concierge a different starting point from a general assistant. Rather than learning only from what a user types in a chat, Zest says it can begin from where someone has actually eaten. For users, the trade-off is straightforward: more personal recommendations depend on sharing transaction-based dining context with the service.

Zest was founded in November 2024 and raised $1.8 million in pre-seed funding from Alexis Ohanian at 776 and Steve Jang at Kindred Ventures, according to TechCrunch.

Booking moves the agent from advice to action

Restaurant recommendations are not new, and TechCrunch notes that finding open tables has been available in other apps for years. The difference Zest is emphasizing is the ability to discuss options and availability in a text exchange, then complete the booking from there.

According to the report, Zest can check availability and make reservations through connected services including OpenTable, Resy, and SevenRooms. Gomez-Hall said Zest books tables through OpenTable or Resy today, with food ordering planned later.

For small businesses and creators, this is part of a wider pattern: AI tools are moving from content generation into task completion. The same shift is showing up in areas such as shopping, browser workflows, and connected apps. Recent examples include Walmart opening an Agent Gateway for AI shopping agents and reports that websites are pushing back on booking and shopping agents.

For Zest, booking is also a business model. Gomez-Hall told TechCrunch that booking and ordering are the startup’s first revenue lines, and that transactions create opportunities for Zest to participate.

What to watch next

Zest began testing the concierge last month, but TechCrunch reports that the company waited to launch until booking capabilities were added. Over time, the startup is considering expanding beyond restaurants into recommendations for businesses, shopping, and more.

That expansion would put Zest closer to a broader personal-agent category, where the assistant is expected to know preferences, negotiate options, and take action across services. We have seen similar positioning around personal AI assistants, including Wabi’s recent personal agent push.

For now, the most practical question is reliability. TechCrunch said its tests worked fairly well, with suggestions that matched taste and constraints such as time of day, party size, and location. But restaurant booking involves live availability, user preferences, and third-party services, so the user experience will depend on how consistently the concierge can turn a casual text into a confirmed reservation.