Restaurant discovery app Zest has launched an AI “restaurant concierge” that lives in your text messages. It recommends places to eat and books the table without you opening an app. TechCrunch AI reports that the startup announced the feature on Thursday. It puts Zest alongside text-based AI agents like Instinct and Poke.
Founder and CEO Mario Gomez-Hall is blunt about the change: “The agent is the product now.”
What Zest actually launched
Zest launched earlier this year with a twist on food recommendations. It connects to your credit cards through Plaid, the financial data company that banks and budgeting apps rely on. That way it can see where you actually eat. According to TechCrunch AI, Zest only looks at food and drink transactions, and it uses them to build a personal map and lists of places you might like.
The new concierge takes that taste profile into a text thread. You text a dedicated number, ask for something to eat, and Zest answers based on your dining history and your friends’.
5 things the concierge can do
- Book tables for you. Zest checks availability and makes reservations through OpenTable, Resy, and SevenRooms. The company had been testing the concierge since last month but held off on launching until booking worked.
- Plan dinner for a group chat. You can add Zest to a group thread. It blends everyone’s tastes, runs an in-thread vote when nobody can decide, and books for the whole group right there in the chat.
- Filter by what matters. You can ask for picks by cuisine, craving, occasion, dietary restrictions, or neighborhood.
- Answer practical questions. It’ll tell you how busy a place might be or whether it has Wi-Fi, and it can send you the menu.
- Skip the getting-to-know-you phase. Gomez-Hall says the agent “doesn’t need weeks of conversation to learn you because it starts from where you’ve actually been.” That’s the main thing setting it apart from general-purpose assistants.
How it held up in testing
TechCrunch AI says the agent “worked fairly well” in its tests. It suggested new places that matched the tester’s tastes and fit constraints like time of day, party size, and location.
TechCrunch also points out that finding an open table isn’t new. Apps have done it for years. The real draw is convenience: you text back and forth about options and get booking times without hopping between apps.
Why this matters
What stands out is the bet behind it. Zest is betting that apps won’t stay the main way people get things done, and that a chat thread will take their place. You can see the same thing happening across the industry, with conversational agents taking over jobs that used to need their own app.
The credit card angle is a real advantage. Most AI assistants learn your preferences by asking questions or reading your chat history. Zest starts with a record of where you’ve spent money. That’s a stronger signal, and it’s also a much more sensitive one. Some people will love the instant personalization. Others won’t want an app reading their card statements, even just the restaurant charges.
Availability and the business model
Right now the concierge covers restaurants only. TechCrunch’s report doesn’t say what it costs. Zest started in November 2024 with $1.8 million in pre-seed funding from Alexis Ohanian’s 776 and Steve Jang’s Kindred Ventures, and it currently has a few thousand users. The company hopes moving to texts will help it grow past that and start making money.
Gomez-Hall laid out the revenue plan: “Booking and ordering are the first revenue lines.” Reservations work today, and food ordering is coming next. In short, Zest wants a cut of the transactions its agent makes happen.
Caveats worth noting
- Small user base. Zest has thousands of users, not millions, so its recommendations depend on a fairly small network of friends’ data.
- Privacy trade-off. You have to connect your bank cards to get the full personalization.
- Familiar features. Reservation booking already exists elsewhere. The difference is the chat interface, not the booking itself.
What comes next
Zest is already thinking past dinner. The startup is considering recommendations for other businesses and for shopping. Gomez-Hall argues that knowing your taste from real spending history is “worth something well beyond dinner.” If he’s right, the restaurant concierge is just the first test of a bigger idea: a recommendation agent built on what you buy, not what you say you like. You can find more details in the original TechCrunch AI report.