The e-commerce platform Meesho is actively integrating artificial intelligence (AI) technologies into all its operational processes. Thanks to its AI-powered personalized ranking system called Prism, nearly seventy percent of all orders are placed through this system. Furthermore, the AI-based customer support agent, Chorus, is currently handling sixty percent of user inquiries.
Pratik Kumar, the company's head of engineering, told Business Standard that AI is a powerful innovation that has allowed them to solve many previously complex problems. He specified that about seventy percent of orders come through the personalized ranking model implemented in the Prism system.
The Prism system uses personalized ranking to suggest products to users. Recommendations are generated based on various signals, including demographic data, geographical location, device usage, browsing history, product clicks, and viewed but ignored ads. According to Kumar, the system can determine buyer preferences even when they visit the platform without a clear intention to make a purchase.
The company also employs the TrendPulse AI system to forecast demand, helping sellers prepare inventory in advance for emerging trends. Kumar cited Navratri as an example, showing how demand can vary depending on the region and day, as buyers look for clothing of specific colors according to local traditions.
According to Kumar, AI is used across a wide range of areas on the platform: from product search and demand forecasting to customer service, voice purchasing, fraud prevention, address prediction, logistics, and seller support.
Meesho is implementing this technology to make the shopping process more accessible for customers who may have difficulty entering search queries or navigating text-heavy interfaces. In such situations, the Vaani AI agent allows users to find products, get product information, add them to the cart, and complete transactions via voice interaction.
For sellers, AI agents help resolve issues related to financial operations, payouts, monetization, and campaign creation, thereby reducing reliance on human involvement and speeding up problem resolution. As the use of AI expands, Meesho is also working to reduce the operating costs of large-scale AI models. Kumar noted that the in-house developed Bharat-MLStack helped reduce inference and model deployment costs by approximately 60–70 percent compared to external providers.
Currently, there are about 960,000 transacting sellers registered on the Meesho platform, and the annual base of such sellers has grown by 87 percent compared to the previous year. AI is applied here to simplify the seller registration process, especially for those located in second and third-tier cities.