India’s premier online fashion retailer, Myntra, has debuted an AI shopping assistant in partnership with Microsoft, named ‘MyFashionGPT’, to provide an AI-powered shopping experience. Departing from conventional keyword-based searches, this technology harnesses the capabilities of generative AI, with a more intuitive, natural language approach.
This integrated AI assistant within the Myntra platform represents a shift in user interaction with online shopping. It was conceptualised to tackle the limitations of conventional search methods. Unlike keyword-based searches, prone to challenges with ambiguous queries, Myntra’s AI assistant adeptly deciphers open-ended requests.
Myntra’s team of engineers explored Microsoft’s Azure OpenAI Service, utilising ChatGPT, a generative AI service, during an intensive two-day hackathon. They discovered the AI’s proficiency in understanding natural language prompts, extending beyond Myntra’s existing product catalogue’s limitations.
Arit Mondal, Myntra’s director of product management, calls this technology ‘big,’ emphasising its significance as the first fashion app to resolve the long-standing ‘search’ challenges in the fashion and lifestyle domain.
Since its launch, the AI assistant has gained rapid traction, significantly influencing user engagement and purchasing behaviors. Users engaging with the assistant exhibited a threefold increase in completing purchases, benefiting from its capability to curate entire looks spanning diverse product categories.
Myntra’s future roadmap includes ambitious plans to introduce features enabling users to visualise complete outfits across various product segments. Additionally, they aim to implement voice search and personalised recommendations, emphasising their commitment to leveraging generative AI to elevate customer experiences continuously.
This integration of technology represents a milestone for Myntra and the wider fashion retail landscape, redefining the online shopping journey and establishing new benchmarks for intuitive, user-centric interfaces.
