Myntra Unveils AI-Powered Shopping Assistant for a Seamless Holiday Experience

An Indian fashion e-commerce platform, Myntra, is leveraging advanced AI technology generative AI to transform how customers search for items, resulting in a more intuitive and tailored shopping journey.

The retail brand has launched an AI assistant that employs ChatGPT the generative AI service offered through Azure OpenAI, which transcends traditional keyword-based searches by comprehending open-ended fashion inquiries.

For example, when a customer faced the dilemma of selecting gym apparel, they turned to Myntra’s AI assistant. With a simple request like \"I need gym wear,\" the system provided a surprisingly precise array of choices, including moisture-wicking jerseys, stylish compression tees, flexible joggers, performance footwear, and fitness trackers, perfectly suited for someone new to working out.

"Deploying Azure OpenAI Service was as straightforward as setting up any server, which gave us a secure method for enabling generative AI within our systems,\" shared Vindhya Priya Shanmugam, Myntra's director of engineering.

Since its launch, the AI shopping assistant This advancement has not only expanded the range of search queries but also significantly boosted the chances of users completing a purchase.

According to Arit Mondal, product management director at Myntra, users interacting with the AI assistant are three times more likely to convert into buyers, as it encourages them to discover and assemble complete fashion looks across various categories.

Addressing Customer Shopping Needs

Myntra recognized the necessity for an AI solution after realizing that conventional search methods, which heavily relied on specific keywords and filters, were inadequate for customers with vague or exploratory questions. In response, during a February hackathon, Myntra’s engineering team explored the potential of Azure OpenAI Service and ChatGPT to see how generative AI could enhance user experience.

The engineers were intrigued by ChatGPT’s proficiency in processing everyday language prompts. By presenting prompts related to fashion inquiries, they discovered that ChatGPT could grasp and address abstract fashion questions, moving beyond the limitations of Myntra’s existing product database.

Over the course of two days, the team tested various prompts to push the boundaries of what generative AI could achieve. Impressively, the AI managed to respond accurately to queries regarding specific outfits, even those absent from Myntra’s catalog, highlighting its potential to solve the perennial \"search dilemma\" in fashion, beauty, and lifestyle.

Myntra credits this outcome to the development of a new feature that allows users to conduct searches using casual language. Utilizing Azure OpenAI Service, the team enjoyed the flexibility to connect various large language models, optimizing for the one that best suited their objectives. This hands-on experimentation enabled them to prototype a functional model in just 48 hours.

Post-hackathon, multiple teams across engineering and product development collaborated to refine both the backend mechanics and the user interface of the AI shopping assistant. This innovative system was officially launched on the Myntra app in late May, timed with one of their flagship events, the End of Reason Sale (EORS).

Looking to the future, Myntra has plans to roll out more features based on generative AI . One feature in the works will allow users to see different product categories together, giving them a holistic view of how items contribute to an overall outfit.

Furthermore, the company aims to introduce voice search capabilities and explore additional avenues for incorporating generative AI into customer support. As Myntra forges ahead, AI is poised to significantly influence the evolution of online fashion retail.