Recommender built by Most Recent.AI
DoorDash has announced Zesty, a social app that leverages AI to introduce users to new restaurants using conversational recommendations as well as community-generated content. “Launching first in the San Francisco Bay Area and New York, the app marks our eXpansion from delivery to a full-scale logistics company by creating solutions at every point of connection between customer and merchant,” DoorDash writes of its new product that goes beyond delivery into social discovery with faster, more personal recommendations than sifting through star ratings, skimming scattered reviews, and parsing influencer videos. Bloomberg first reported news of the launch.
How Zesty’s AI restaurant finder and social app works
Using a DoorDash account, customers start by chatting with an AI bot, which will make personalized recommendations — everything from “cozy pasta spots for introverts in Williamsburg” to “brunch that’s good for groups.” The system even has suggestions and the ability to understand context, refining its results as it learns your interests, neighborhoods, cuisines, and budget comfort zones.
Recommendations are intended to be actionable. You can save places to lists, share them with friends, and follow people whose taste you trust. A familiar social network, Zesty includes diners’ photographs and comments, surfacing real-world experiences next to AI summaries. Andy Fang, a co-founder of DoorDash who is also one of the founders behind X, explained on X that Zesty collects signals from DoorDash, Google Maps, TikTok, and other places to present recommendations in an effort to cut down on time spent researching as well as decision fatigue.
Why DoorDash Is Betting On Social Discovery
Americans’ spending on dining is gargantuan, with the National Restaurant Association in January projecting industry sales would top $1 trillion this year. Meanwhile, discovery is now happening on feeds and creators. An executive at Google has observed that nearly 40% of young users now begin their searches for places to eat — a rich source of commercial intent, in marketing lingo — on TikTok or Instagram. That behavior shift carves out space for a purpose-built, AI-first experience focused specifically on dining — not just general search.
DoorDash can provide that scale — tens of millions of consumers, hundreds of thousands of restaurant partners — to seed a network effect. It has also tested features for in-person reservations and in-store rewards, indicating a broad strategy that includes delivery, pickup, and dine-in. Zesty matches that arc to capture intent at the top of the funnel, and then convert discovery into reservations, walk-ins, or delivery orders.
The crowded field of social dining discovery platforms
Zesty enters a competitive landscape. Google Maps now has more than a billion monthly users and is the default for millions. Yelp has more than a million reviews and photos — taken together, in the hundreds of millions; TikTok and Instagram create viral food trends that can make or break a hole-in-the-wall spot overnight. The tricky thing for a new entry is trust: That is, can AI recommend restaurants that feel credible, contemporary, and true to one’s taste without overwhelming them?
DoorDash’s difference from everyone else might be intent plus transaction. If a user finds out about a ramen bar on Zesty, DoorDash might help that person get the food ordered on a rainy night or make a reservation for their birthday — cutting out drop-off between suggestion and action. That closed loop, together with the social graph it has (your friends, creators, and local experts), could chip away at the “too many tabs” problem we all suffer from today as part of restaurant research.
Data quality, privacy, and accuracy concerns with AI
Generative systems can be prone to hallucinations and outdated information. Zesty’s approach — pooling across platforms and layering on user photos and comments — is designed to ground outputs of AI in verifiable, current matter. You’d expect guardrails around claims, such as hours, pricing, and availability — all areas where traditional review sites (and mapmakers) still have an edge through structured data.
Privacy will be another test. Location and preference information can make recommendations more acute, but it has also come under scrutiny. Because regulators are keen to see data minimization and transparency, Zesty’s onboarding and settings will have to more clearly explain how it’s using signals, what it shares socially by default, and the opt-out for personalization.
Monetization plans for Zesty and key metrics to watch
In the short term, Zesty’s value is in engagement: time spent finding and saving stuff, and sharing. Monetization over time may include sponsored placements, creator partnerships, affiliate fees on reservations, and higher order conversion for DoorDash’s marketplace. The metrics to watch for in the Zesty test will be recommendation quality (saves and shares per session), creator adoption, and how quickly Zesty is able to convert actual visits or sales.
If Zesty persuades enough diners that one chat can replace the rummaging and juggling of reviews, reels, and maps, door-to-door meals could have a lasting say in where and how people choose to eat. Testing in two food-obsessed markets out of the gate is a smart stress test; whether AI-driven social discovery can scale to those who haven’t yet adopted isn’t clear.