Meesho’s Voicebot Cuts Call Costs by 75%

Last month, Bengaluru-based e-commerce platform Meesho deployed an AI-driven voicebot that manages about 60,000 calls per day, cutting customer call costs by 75%. 

“We have built India’s first GenAI voicebot, which today operates in two languages, Hindi and English,” said Sanjeev Barnwal, co-founder and CTO of Meesho. “Today, it’s operating at about 60,000 calls per day, so this is where we believe that we are the first ones to kind of launch at this scale in India.”

On cost savings, Barnwal noted, “On the voice front, we see cost per call, which is now 25%. It has gone down by 75%.” He added, “When we continue to scale it, I suspect the cost will go down from 25% to maybe 15%.”

Addressing user preferences, Barnwal said: “Our user base is targeted towards those who prefer affordability and accessibility on a wide selection of products. This is a value we bring to the table.” He stated that early gains from AI tools have not only reduced costs but also supported the company’s move toward profitability.

Meesho plans to maintain its momentum with ongoing efforts to handle more complex calls through automation and potentially add new languages, ensuring greater accessibility and convenience for customers across India.

Meesho and its LLMs

Meesho is also adopting LLMs to scale other parts of its operations, from customer support to product cataloguing, while choosing Azure OpenAI for sustainable cost optimisation and GitHub Copilot for its more than 200 developers.

“There are many teams within Meesho that are successfully deploying LLMs in their applications,” said Debdoot Mukherjee, chief data scientist and head of AI at Meesho, in an interview with AIM. For instance, in customer support, when dealing with large volumes of inquiries, it becomes challenging to hire enough staff to manage and scale that workload effectively.

Over the last financial year, Meesho’s revenue grew by 33% to ₹7,615 crore. The company credits this growth partly to AI-driven efficiencies. 

“This success stems from driving efficiencies across multiple areas like logistics, as well as leveraging generative AI and machine learning for better discovery, improved in-app experience and round-the-clock customer support,” Meesho’s team stated.

Meesho is Not Alone in Embracing GenAI 

Other Indian e-commerce leaders like Flipkart, Amazon, and Myntra have also launched AI features, such as search assistants and personal style guides. Flipkart’s senior product manager Aradhya Saxena said that their AI-powered search feature, Immerse, aims to replicate how Indian shoppers interact with storekeepers, refining searches until users find exactly what they want.

Amazon introduced Rufus, a shopping assistant powered by generative AI, while Myntra launched My Stylist and MyFashionGPT to give fashion recommendations and simplify online searching. 

Nykaa and Lenskart, along with others, have followed suit, adopting similar technologies to stay competitive in India’s evolving online shopping landscape.