Delhi, India, 25th March 2026 - With more than 25 million active monthly users—and hundreds of millions of food delivery orders annually —Indian-born Zomato is the world's second-largest food delivery company. At the heart of the business’s success is Zomato’s ability to seamlessly scale, manage complex data, and build innovative AI powered applications at pace.
MongoDB Atlas, a fully managed data platform, has made this possible. In 2017, Zomato moved all of its self-hosted clusters to MongoDB Atlas. The company has expanded greatly since then, and Zomato now uses MongoDB Atlas to power a number of their key operational systems, including customer experience.
Particularly, Zomato built Nugget, an AI-native CX platform engineered to manage high-volume customer support through autonomous, multi-channel AI agents. First used in-house for Zomato’s own customer service needs, Nugget was scaled into a centralized platform and launched as an enterprise offering in 2025, making it available to a wide range of industries.
Unlike traditional bots or ticketing systems, Nugget orchestrates voice, chat, social, and in-app interactions through a unified operational layer that integrates directly with backend systems.
Built on MongoDB Atlas, Nugget leverages real-time context storage, scalable data pipelines, and intelligent workflow execution to automate complex customer journeys end-to-end. This architecture has enabled Zomato to reduce support costs by 55%, while preserving service quality and operational resilience at scale.
Over the years, Zomato’s support costs had grown to US$20M annually: operations were spread across 10 disparate platforms, resulting in fragmented data and limited insight. While approximately 60% of customer support tickets were auto-resolved, Zomato wanted to further improve customer satisfaction scores.
These challenges compelled Zomato to overhaul its approach to customer experience, and to build a bespoke in-house, AI-native platform: Nugget.
After a rigorous testing process in which databases including DynamoDB and DocumentDB were evaluated, Zomato decided to expand its use of MongoDB Atlas, and use it as the data layer for Nugget.
Zomato tested each solution for performance, reliability and scalability. MongoDB Atlas’s flexible schema and high performance helped it rank first across all criteria. Furthermore, MongoDB Atlas's aggregation pipeline helped with dynamic querying capabilities, as well as providing strong write consistency and compound indexing which further increased performance. Finally, MongoDB Atlas’s built-in encryption features ensure all sensitive customer data is encrypted in the database. This assists Zomato’s ability to comply with stringent security requirements, while efficiently managing the growing volume of historical data.
Powered by MongoDB, AI agents within Nugget can independently take actions based on real-time context and policy logic. The platform also includes an agent co-pilot that assists human agents with contextual insights, recommended actions, and historical data, alongside a unified omnichannel workspace.
Nugget can store evolving conversational context, ticket metadata, and user history in a flexible data model. This enables AI agents to reason across fragmented signals and route issues dynamically. This foundation allows Nugget to automate triaging, track outcomes across workflows, and surface actionable insights through advanced analytics.
“Nugget was built to reflect the reality of enterprise customer support, where conversations are fragmented, context is constantly evolving, and automation must coexist with human judgment,” said Vivek Gupta, Senior Developer, Nugget, at Zomato. “MongoDB Atlas provides the flexibility and performance we need to power real-time AI workflows while maintaining reliability and scale across industries.
As of January 2026, Nugget handles more than 15 million conversations per month, achieves an average query resolution rate of 85%, and improves human agent productivity by 40%. It has also helped reduce customer service costs from $20 million down to $9 million annually.
While Nugget was initially developed to address Zomato’s own scale and operational complexity, it was architected as a configurable, multi-tenant platform from the outset. This design has enabled Zomato to scale Nugget into a centralized, enterprise-grade platform, which was launched as an enterprise offering in 2025.
As of 2026, Nugget is used across industries such as financial services, hospitality, and healthcare, where high ticket volumes, strict compliance requirements, and real-time decision-making are essential.
MongoDB Atlas’s multi-tenant architecture enables Nugget to onboard new enterprise customers while maintaining strict data isolation: each customer can operate on a dedicated database instance aligned to their deployment and compliance requirements—be it on the cloud of their choice, or on-premises.
Zomato continues to build on MongoDB, expanding Nugget with additional AI-driven capabilities, with a focus on increasing autonomous resolution rates, improving agent collaboration, and enabling faster, more consistent customer experiences.