75% of Meesho orders are now driven by AI recommendations, says CEO Vidit Aatrey
text_fields
Meesho is increasingly adopting an AI-first approach as it looks to strengthen its position in India’s e-commerce market, founder and CEO Vidit Aatrey said.
Aatrey said more than 70 per cent of the company’s code is now generated using artificial intelligence, while the number of platform experiments has doubled compared to last year.
He said the biggest transformation is taking place in how customers shop on the platform. More than 75 per cent of orders now come through personalised recommendation feeds instead of direct search, powered by Meesho’s in-house AI system called PRISM, short for Personalised Ranking and Intent Signal Module.
According to Aatrey, the shift represents a move from traditional intent-based shopping to AI-led discovery, where the platform predicts user preferences before they are explicitly searched for.
The company has also developed several AI tools aimed at solving India-specific challenges. These include Vaani, designed to help first-time internet users complete purchases, GeoIndia LLM, a location intelligence model built for Indian addresses, and TrustMesh, a risk engine that blocked 9 million high-risk transactions in FY26.
Aatrey said every customer interaction helps improve the platform’s AI systems further.
The company’s financial performance also showed strong growth. Meesho reported a 43 per cent year-on-year rise in fourth-quarter Net Merchandise Value to Rs 11,371 crore. Orders also increased 43 per cent to 717 million, while annual transacting users rose 33 per cent to 264 million in FY26.
The company said losses narrowed by around 66 per cent due to operating leverage and AI-driven efficiencies.
Aatrey said Meesho’s long-term focus remains on making online commerce accessible to underserved users, including homemakers, small shopkeepers and manufacturers with limited distribution access.