Choosing a restaurant can mean jumping between recommendations, maps and booking apps. Zest is bringing those steps into a text conversation, where its AI concierge can suggest places based on dining history and help book a table.
Recommendations built from dining history
Zest launched earlier this year as a restaurant discovery app. It links with users’ credit cards through Plaid to identify food and drink transactions, which the company uses to build personalized maps and lists. Zest says it focuses on those categories when using the transaction data.
That history is meant to give the concierge a starting point for understanding a person’s tastes. Rather than relying only on preferences shared during a conversation, it can draw on places the user has actually visited, as well as friends’ histories.
Users can request recommendations by cuisine, craving, occasion, dietary restrictions or neighborhood. They can also ask questions about restaurants, including how busy a place may be or whether it offers Wi-Fi, and ask Zest to send a menu.
From a text suggestion to a reservation
The new feature moves Zest’s restaurant recommendations into text messages. Users contact the concierge at (212) 470-6067, where they can discuss options and ask about availability. Zest can then make a reservation through connected services including OpenTable, Resy and SevenRooms.
The company says the booking capability was a key reason it waited to launch. It began testing the concierge last month, but held the release until it could add reservations. In TechCrunch’s tests, recommendations reflected factors such as time of day, party size and location, while the back-and-forth made it convenient to compare available booking times.
Finding open tables is already possible through other apps. Zest’s approach combines that task with restaurant discovery in a single conversation, letting users move from “where should we eat?” to a booking without leaving the thread.
Planning dinner with a group
The concierge can also join a group chat. It can take into account the tastes of everyone in the conversation and suggest places that could work for the group. If people are stuck choosing, an in-thread voting feature can help settle the decision.
Once the group picks a restaurant, Zest can book a table without requiring participants to switch to another conversation or app. The group chat becomes a shared place to discuss options, vote and arrange the reservation.
A wider role for the agent
Zest’s launch reflects a broader idea: conversational agents can handle tasks that once required people to open separate mobile apps. The company is positioning the concierge as its product interface, with restaurant suggestions, questions and bookings handled over text.
For now, the service is focused on restaurants. Zest is considering expanding into recommendations for businesses, shopping and other areas. The company also sees booking and, later, food ordering as potential revenue sources, with transactions handled through connected services.
That expansion depends on whether the personal context that informs restaurant suggestions can be useful for other choices, too. The company’s bet is that a service informed by where people have eaten can make recommendations more relevant, while text makes it easier to act on them.