Deepinder Goyal founded Zomato as a restaurant discovery platform. Fifteen years later, he runs what is, by any reasonable measure, one of the most data-rich companies in India — a company that knows what 100 million people eat, when they eat it, where they order from, and how much they are willing to pay. The question he has been wrestling with for the past two years is how to turn that data advantage into an AI advantage.
The Data Foundation
"We have been an AI company for years without calling ourselves one," Goyal says. "Every recommendation we make, every delivery route we optimise, every price we set — these are AI decisions. What has changed is the sophistication of the AI and the ambition of what we are trying to do with it." Zomato's data assets are genuinely extraordinary. The company has detailed records of hundreds of millions of food orders, including not just what was ordered but when, from where, in what weather, on what occasion, and how the customer rated the experience.
Demand Forecasting at Scale
The most economically significant AI application at Zomato is demand forecasting — predicting, with high accuracy, how many orders will be placed in each area at each time of day. Zomato's demand forecasting models now predict demand with accuracy that allows the company to pre-position delivery partners in areas where orders are likely to surge before the surge actually happens. "The difference between a good demand forecast and a bad one is not just efficiency," Goyal says. "It is the difference between a delivery partner who earns a good income and one who spends half their time waiting for orders."
Personalisation Beyond Recommendations
Zomato is now deploying personalised pricing and promotions — AI systems that identify each customer's price sensitivity and preferences, and tailor offers accordingly. "We used to send the same promotion to everyone," Goyal says. "Now we send different promotions to different people based on what we know about their behaviour. The result is that we spend less on promotions and get better outcomes."
The Customer Service Transformation
Zomato handles millions of customer service interactions every month. The company has deployed AI systems that handle the majority of routine interactions automatically, escalating to human agents only when the situation requires human judgment. The human agents who remain are now focused on the complex, emotionally sensitive interactions where human judgment genuinely matters. Their job satisfaction has improved, their productivity has increased, and the overall quality of customer service has gone up even as the cost has gone down.
What Becoming an AI Company Actually Means
"In five years, the food delivery companies that survive will be the ones that are best at AI," Goyal says. "Not the ones with the most restaurants or the most delivery partners or the most marketing budget. The ones that are best at predicting what customers want, optimising how it gets to them, and making the whole experience feel effortless. That is an AI problem." The investment Zomato is making in AI infrastructure is, in Goyal's view, the most important investment the company is making — one that shows up not in product announcements but in the small improvements in delivery time, recommendation quality, and customer satisfaction that compound over time into a durable competitive advantage.