# Zest AI Restaurant Concierge: Recommendations and Reservations Over Text

Zest brings restaurant discovery into messaging, with optional dining-history data that makes its privacy choices worth understanding.

By Paras Madan | 2026-10-31

Zest's new AI restaurant concierge lets people ask for recommendations and arrange reservations over text. It works without installing an app, while an optional card connection can personalize suggestions using dining history. The interesting question is whether that extra context makes choosing a restaurant meaningfully easier.

[TechCrunch reported the launch](https://techcrunch.com/2026/10/30/zests-new-ai-agent-finds-restaurants-and-books-reservations-over-text/) on October 30, following its introduction the day before.

## How Zest's text concierge works

According to [Zest's product page](https://www.textzest.com/), users can message from an iPhone or Android phone and describe what they want, including the neighborhood, cuisine or occasion. The service can suggest places, coordinate preferences for a group and help with reservations.

Zest currently describes the concierge as free, without subscription or paid features. Connecting a card is optional, and the service still works without it. It also says it may proactively send suggestions roughly weekly; users can opt out of messages with STOP.

The company's privacy notice says booking requires a connected reservation account. A conversational request therefore should not be confused with a confirmed table. The useful outcome is the completed reservation, not simply a promising recommendation.

## What connecting a card actually shares

Zest's [privacy policy](https://www.zestmaps.com/privacy), updated October 28, describes a Plaid connection for transaction data. Zest says it does not receive banking credentials or full account and card numbers.

The policy says it filters transaction records to food and drink activity. Relevant fields include the merchant, date, location and amount. An initial connection can cover approximately two years of history, followed by updates typically once a day.

That creates an important distinction: **not displaying spending amounts is different from not receiving them.** The policy says amounts are not shown in the product or used to infer income or creditworthiness. Disconnecting stops future collection.

The messaging service also processes conversation content, including attachments where provided, through messaging infrastructure and AI providers. For group suggestions, Zest says participants see recommendations rather than the underlying visit histories used to produce them.

## Why dining history could matter more than the chat interface

A restaurant request often contains less information than the person making it assumes. “Somewhere good for dinner” leaves out familiar places, price expectations, dietary preferences and how far the group will actually travel.

Past visits can help narrow that gap, but they are imperfect evidence. A transaction might represent a work lunch rather than a favorite restaurant. A repeat visit might reflect convenience rather than enthusiasm. The product still needs a way to accept corrections instead of treating every purchase as a preference.

The group case is harder again. Finding somewhere one person likes is different from finding a reasonable compromise for several people. If the concierge reduces that back-and-forth without exposing everyone's history, it could remove more friction than a better ranked list of restaurants would.

## What remains unproven

The public product description does not establish booking success rates, how often recommendations fit a group's preferences or how reliably the service handles a failed reservation. Those are the results that would tell us whether the convenience survives ordinary use.

For now, the most useful distinction is between a restaurant chatbot and a working concierge. The latter needs to carry a preference through discovery, availability and confirmation, while making clear when the user still needs to act. That is where Zest's experience will have to earn its place in someone's messages.

I'm Paras. I work at the intersection of AI, engineering and distribution.

I help technical founders turn product behavior and AI launches into clear stories that people understand and share.

[Book a 30-minute call](https://cal.com/paras-madan/30min).