Swiggy Opens Commerce Services to External AI Agents
Food and grocery delivery platform Swiggy is integrating its services with external AI agents, allowing users to order food and book tables beyond its app.
Indian technology company Swiggy is expanding the accessibility of its commerce services by integrating them with external AI agents. This enables users to discover food, build carts, and book restaurant tables outside the Swiggy app, utilizing AI assistants.
Swiggy utilizes four Model Context Protocol (MCP) servers that act as connectors, allowing AI tools to interact with its services. These currently span food discovery, Instamart grocery deliveries, Dineout restaurant reservations, and Scenes event services.
By allowing external agents to access its services, Swiggy aims to reach customers even when their shopping journey doesn't start within its app. CTO Madhusudhan Rao explained how users can connect the MCP to their preferred large language models to request tasks such as meal planning based on dietary needs.
The company is also refining its internal AI systems to reduce reliance on any single model. Swiggy has implemented an LLM gateway that directs tasks to different models. Additionally, it uses models from multiple cloud providers, optimizing choices based on availability, response time, and cost.