As global tech giants race to reinvent online shopping through AI, Meesho believes the future of e-commerce in India will not be built around search bars and product listings. Instead, it will look more like a conversation with a neighbourhood shopkeeper.
Speaking to AIM, Debdoot Mukherjee, Chief Data Scientist and Head of AI and Demand Engineering at Meesho, says the social commerce company is using AI to fundamentally rethink how users discover, evaluate, and purchase products online.
“The user experience and interfaces for shopping have been due for a rethink for a long time,” Mukherjee says. “You start with a product listing page, move to a product details page, add items to a cart, and finally check out. It is inspired by how supermarkets work in the US.”
According to him, that model does not naturally fit the way Indians shop.
“When we walk into a store in India, the shopkeeper has already assessed what kind of customer we might be. Shopping is conversational. The shopkeeper recommends products, answers questions, and helps us make decisions. E-commerce today feels lonely because the shopper is on their own.”
The AI Engine Behind 75% of Orders
Meesho’s proprietary AI system, PRISM (Personalised Ranking & Intent Signal Module), now powers more than 75% of orders on the platform. The company says the system understands user intent in real time.
“Pretty much every surface of the Meesho app today is powered by PRISM,” Mukherjee explains. “When you land on the homepage, when you browse products, when you search, PRISM is deciding which products are most relevant for you.”
According to the company, PRISM continuously analyses behavioural, transactional and contextual signals across users and browsing journeys through a network of more than 100 AI ranking models. The system is trained on over 400 trillion input signals and executes more than 6 trillion inferences daily within milliseconds.
Mukherjee credits Meesho’s AI progress to years of investment that started well before generative AI became mainstream.
“We started our AI journey nearly five years ago, much before the recent AI boom,” he says.
The company has steadily evolved from smaller machine learning models to increasingly sophisticated AI systems capable of deep personalisation. The BharatMLStack, Meesho’s cloud-agnostic machine learning infrastructure that was open-sourced last year, serves as the foundation on which many of these systems operate.
Solving India’s Typing Problem
For Meesho, AI is not simply about improving recommendations. It is also about making commerce accessible to users who may struggle with traditional digital interfaces.
Mukherjee calls it India's ‘typing problem.’
“People think in their mother tongue and express themselves using a Latin keyboard,” he says. “The result is often highly ambiguous and noisy.”
This creates challenges for conventional search systems that rely heavily on keywords.
AI allows Meesho to understand user intent even when searches are incomplete, misspelt, or written in vernacular languages. “PRISM plays a big role in understanding search keywords expressed in vernacular language and delivering relevant results,” he says.
PRISM supports more than 10 Indian languages, including Hindi, Bengali, Marathi, Tamil, Telugu, Kannada, Malayalam, Gujarati, Punjabi, and Odia.
The company previously also launched Vaani, its AI-powered voice shopping assistant that allows users to discover products, ask questions, compare options, and even complete transactions through natural conversations.
The company is also working to support users who may not be comfortable reading reviews, interpreting ratings, or comparing products. Meesho’s AI can simplify these complexities by helping users understand products through natural language and voice interactions.
Meesho said PRISM's capabilities include personalised recommendations, trend intelligence, catalogue understanding, and seller growth tools. Through Trendpulse, an LLM-powered discovery engine, the system identifies emerging demand patterns across regions and cities and surfaces products based on changing consumer preferences.
Competing in a World of AI Commerce
The timing of Meesho’s AI push comes as Big Tech companies aggressively enter commerce.
At Google I/O 2026, Google unveiled Universal Cart, a shopping experience built to work across products and services. Meanwhile, OpenAI has introduced shopping capabilities within ChatGPT, allowing users to discover and purchase products directly through conversational interfaces.
Despite the competition, Mukherjee believes success in Indian commerce will depend less on AI models and more on understanding local users.
“The player that truly understands the user and comes up with innovations that are user-first will win in this market,” he says. “Our mission is to unlock and democratise internet commerce for a billion users.”
Meesho made its stock market debut in December last year. Since listing, its shares have jumped nearly 2%, hovering at ₹168 apiece.
- OpenAI Says AGI is Here. But Not Everyone is Buying It
- Why Oracle is Bringing Post-Quantum Security to Java
- What Really Happens When You Turn Your Selfie Into a 1980s AI Pic?
- AI Agents Are Making Data Security A Bigger Enterprise Problem
- OpenAI is Getting Nervous About Reinforcement Learning
- Meta’s Bold Open-Weight Bet is Meant to Muse Developers