I stood in the middle of a crowded food court in Bengaluru last Tuesday, watching a group of six engineers argue about lunch for 45 minutes. It’s a scene I’ve seen hundreds of times, in New York and London and just about any major tech hub on the planet. The debate wasn’t about the food. It was about the friction of agreeing.
Who likes spicy? Who is avoiding dairy? Who has dietary restrictions that nobody else remembers? And who wants to be the person to actually pick up the phone and make the reservation?
That’s where Zest comes in. Or rather, that’s where Zest thinks it comes in. The new AI agent doesn’t just recommend a place. It runs the entire social negotiation process over text.
This isn’t a search engine with a chatbot wrapper. It’s a concierge that votes, compromises, and books, all before your group has finished deciding who is paying for the appetizers.
The Death of the “Let’s Just Pick a Place” Argument
Most food apps are passive. You search, you filter, you scroll. You are the user. Zest flips that model on its head.
The AI initiates the conversation. It sends a poll. It asks questions specific to the group dynamic, not just the user. If one person is a strict vegan and another is a carnivore with a shellfish allergy, the system identifies the overlap immediately.
It’s a surprisingly delicate algorithmic dance. The company says the agent can handle up to 12 people in a single thread. That’s a room full of colleagues. Or a large family reunion.
The friction of consensus is the biggest killer of group plans, and Zest is betting that AI can smooth it out without the social awkwardness of forcing a decision. But there’s a catch. Trust. You are letting an algorithm pick where you spend your money and your social capital. If the AI suggests a trendy spot with bad reviews, who is to blame?
The app? The AI? The person who hit “approve”? This is the first real test for consumer AI agents moving from novelty to utility.
It’s not enough to be smart. It has to be right. Or at least, it has to be good enough that no one complains. That’s a harder sell than it sounds.
We’ve seen chatbots give bad medical advice. We’ve seen them write emails that get people fired. Now they are deciding where we eat dinner. The stakes seem lower, but the frequency of use is going to be incredibly high.
Every day. For every meal. The margin for error is thin.
How the “AI Concierge” Actually Works
The mechanics are simple on the surface. You open the app. You tag your friends. The bot wakes up.
It asks each person what they’re in the mood for. It cross-references their historical data—what they ate last week, what they rated highly, what they penalized automatically. Then it proposes three options. The group votes.
If the vote is split, the AI mediates. It might suggest a compromise venue that has a vegetable menu that isn’t sad and a meat section that isn’t iffy. It’s a negotiation engine. Not a search tool.
This distinction matters. Search tools return lists. Negotiation engines return decisions. The goal is closure.
You want to stop talking about food and start eating it. There’s a subtle shift in power here. The human is no longer the driver. The AI is the driver. The human is the passenger.
And the passenger’s job is just to signal comfort or discomfort. If you don’t like the location, you say so. The AI pivots. If you like the price point, you stay.
It’s a feedback loop that happens in real-time, over SMS or WhatsApp, depending on where the users are. This is critical. Most food apps live in their own silos. You have to open the app.
Zest lives in the chat you’re already in. It inserts itself into the existing communication flow rather than demanding you switch contexts. That’s a small UX detail, but it’s likely the thing that makes or breaks adoption. Context switching is the enemy of convenience.
Zest is fighting that war. India is the world’s largest market for food delivery. But it is also a market where “delivery” often struggles with the “dining out” culture. In cities like Bengaluru, Mumbai, and Delhi, the social act of eating together is deeply rooted. It’s not just about nutrition.
What This Means for India’s Tech Hub
It’s about connection. The weekend dinner out is a ritual. And rituals require coordination. For Indian users, the barrier isn’t just finding a good restaurant.
It’s finding a restaurant that can handle 8 people, has a non-AC option because someone hates cold air, and has a price cap that doesn’t require a CFO approval. Zest’s model hits a nerve here. The “group voting” feature is essentially a digital version of the cousin who always asks, “But can we get it spicy?”
Consider the scale. India has over 100 million registered users on food apps. But the share of the market that actually books tables, rather than orders delivery, is still growing. Why?
Because booking is hard. Phone calls fail. Websites are clunky. Apps are fragmented.
Zest solves the booking friction with AI. But there’s a deeper cultural angle. In India, “saving face” is a big deal. If the AI picks a place that is perceived as “too fancy” for the group’s dynamic, it could be awkward.
The AI needs to understand social semiotics, not just flavor profiles. It needs to know that a 5-star Michelin spot might be a mistake for a casual date. It needs to read the room. And the room is a group chat.
That’s a complex dataset to parse. And it’s a huge opportunity for a startup that understands local nuance. The global product might struggle. The localized version could dominate.
The Business Model and the Data Goldmine
Who is paying for this? Zest doesn’t explicitly say it takes a cut of the bill. That’s a suspicious omission. In the food tech world, data is the product.
If Zest knows that a group of 5 engineers in Brooklyn eats at specific Italian places on Fridays, that data is valuable. To the restaurants? To the advertisers? To the insurance companies figuring out risky dining habits?
The potential for monetization is vast. But it’s also invasive. You are sharing your taste, your budget, your social circle, and your location, all in one package. The trade-off is convenience for privacy.
And in 2024, consumers are warier than ever. They don’t want to sell their soul for a faster table reservation. Or do they? Maybe they do.
Maybe the time saved is worth the data leak. That’s a gamble. And it’s a gamble the industry is taking more often.
There’s also the question of restaurant adoption. Restaurants are conservative. They hate apps that give away discounts. They hate apps that bypass their own booking systems.
Zest has to convince restaurant managers that this is a source of revenue, not a disruption. It’s about filling empty tables. It’s about predicting demand. If the AI knows a big group is coming, the restaurant can prep.
That’s efficiency. That’s money. But it requires trust. And trust is built slowly.
One bad reservation, one no-show, one complaint, and the restaurant might pull the plug. The ecosystem is fragile. And it’s all riding on the AI being consistently good. If it’s 90% accurate, it’s a joke.
If it’s 99% accurate, it’s a utility. The line between those two numbers is where the business lives.
A Glimpse of the Future We’re Already Living In
Let’s step back. This isn’t just about food. This is a prototype. Zest is building a “life concierge” agent.
Food is just the entry point. Why? Because food is universal. Everyone eats.
Everyone argues about it. It’s the low-hanging fruit for AI agency. Next up? Travel.
Zest will book the flight, the hotel, the car. The AI will combine everyone’s itineraries and find the overlap. Then it’s events. Concerts.
Sports. The pattern is clear. We are moving from apps that do one thing, to agents that do many things, in sequence. The user stops thinking in apps.
They start thinking in outcomes. “I want to go to Paris.” “I want to eat sushi.” The AI handles the rest. It’s a shift in cognitive load. We are offloading our decision-making to machines.
And we don’t even notice it happening. Because it’s so smooth. So seamless. That’s the danger.
We’re losing the ability to make the small decisions for ourselves. And we’re okay with that. That’s the real story here. Not the AI.
The us. But is that bad? Maybe not. Maybe we’re too busy. Life is fast.
Attention is scarce. If an AI can handle the mundane friction of choosing a restaurant, so we can focus on the human connection over the meal, that’s a win. The food is still the point. The reservation is just the logistics.
And logistics is boring. We like to let robots do the boring stuff. It’s a logical progression. It’s also a slippery slope.
Where does it end? Does the AI start suggesting who we should date? Based on our dining habits? Based on our budget?
Based on our dietary compatibility? That’s a different monster. But it’s the same engine. The implications are vast.
And they’re arriving fast. We don’t have time to hate it. We’re too busy using it.
Frequently Asked Questions
Is Zest free to use?
As of now, Zest appears to be free for users. The business model likely relies on data monetization or commissions from restaurants. Users should watch for any subscription tiers or premium features that might appear as the platform scales. The initial free access is a common strategy to build habit and data sets.
Does Zest work with all restaurants?
Not yet. Zest is building a merchant network. It starts in major metros with high dining density. If your favorite local spot isn’t listed, the AI will likely suggest alternatives.
The network is growing, but it isn’t universal. Check the app for available partners in your specific neighborhood before relying on it for a niche request.
How does the AI handle dietary restrictions?
The system asks each user for their restrictions during the initial setup. It then filters options in real-time. If one person is gluten-free and another is vegan, the AI finds restaurants that can serve both comfortably. It doesn’t just list restaurants.
It validates the menu against the group’s needs. This reduces the risk of awkward “wait, is this gluten-free?” moments at the table.
Can I use Zest for large groups?
Yes. The system is designed for groups up to 12 people. It manages the polling and consensus-building for all members simultaneously. This is particularly useful for corporate dining or large family gatherings.
The AI acts as a project manager for the meal, ensuring everyone’s voice is heard and the logistics are handled. It’s a time-saver for group leaders. Source: TechCrunch
Source: TechCrunch
