Meesho Says 70% of Its Code Is AI-Generated; AI Feeds Drive 75% of Orders

Why It Matters

Meesho’s deep AI integration showcases how emerging markets can leverage generative tools to slash development cycles and drive higher e‑commerce conversion, setting a benchmark for cost‑efficient scaling. The financial impact underscores the trade‑off between rapid growth and short‑term profitability.

Key Takeaways

• 70% of Meesho’s code now generated by AI tools
• AI-powered feeds generate 75% of platform orders
• Conversion up 15% and seller onboarding 27% faster
• Voice assistant Vaani attracted 1.5 M users, 22% higher conversion
• In‑house ML stack cuts infrastructure cost by up to 70%

Pulse Analysis

Meesho’s announcement that more than seven‑tenths of its codebase is now AI‑generated reflects a broader shift in software engineering toward generative models. By automating routine coding, testing, and deployment tasks, the platform has halved iteration times, allowing it to double the number of experiments launched in Q4 FY26. This rapid development cadence mirrors moves by global tech giants, but Meesho’s scale in a price‑sensitive market highlights how AI can democratize high‑velocity product cycles without massive R&D spend.

The impact of AI on Meesho’s consumer experience is even more pronounced. Personalized AI feeds now surface in three‑quarters of all transactions, translating into a 15% lift in conversion and a 27% reduction in the time new sellers need to gain traction. The rollout of Vaani, a voice‑enabled shopping assistant, further deepens engagement, especially among regional‑language users, delivering a 22% conversion bump for early adopters. These data points illustrate how AI‑curated discovery and voice interfaces can reshape buying habits in emerging e‑commerce ecosystems.

On the cost side, Meesho’s proprietary machine‑learning stack delivers 60‑70% lower infrastructure expenses than comparable cloud services, processing over 6.4 trillion inferences at peak load. Coupled with a 20‑point improvement in geocoding accuracy via its Geo‑India model, the AI‑driven logistics layer sharpens delivery efficiency. While the company logged a $164 million loss due to aggressive user‑acquisition, the operational savings and revenue uplift from AI suggest a pathway to profitability as the platform scales its 264 million users and 2.67 billion annual orders.