One of the first criticisms leveled at online apparel sellers was customers’ inability to try products on prior to purchase — and despite many advances in technology, that fact remains a major ecommerce challenge. However, Amazon is using AI to try to close the appearance vs. reality gap in multiple ways.
According to a blog post written by Apoorv Chaudhri, Director of Computer Vision and Machine Learning at Amazon Fashion, AI-powered solutions now offer Amazon shoppers:
- AI-driven size recommendations personalized for each customer;
- Fit Review Highlights using relevant customer feedback; and
- Reimagined size charts drawing on Large Language Models (LLMs) to extract and clean product size chart data from multiple sources.
For brands, Amazon offers a Fit Insights Tool based on LLMs that contextualizes returns and size chart analyses with customer reviews and uses machine learning to identify defects in size charts. With this data, brands can better understand customer fit issues, improve how they communicate sizing to customers and even incorporate the feedback into product designs and manufacturing.
The personalized sizing recommendation feature for shoppers uses an algorithm that anonymously clusters customers with similar size and fit preferences and products with a similar fit. The algorithm learns from millions of product details, including style, size chart and customer reviews, as well as billions of anonymized customer purchases. It also takes into account sizes bought — and kept — by similar customers for the same product or similarly fitting products.
Advertisement
The tool also takes into account that sizes change over time. For example, a customer that bought children’s pants in a certain size this month will likely need a larger size for the next purchase.
The Fit Review Highlights function tells customers whether to size up or down in a particular style, based on reviews from customers who have purchased the item in the same size. It uses advanced AI and LLMs to extract details from customer reviews such as size accuracy, garment fit on specific body areas and fabric stretch.
Amazon also is seeking to make sizing charts more accurate and easier to use, and is leveraging LLMs to automatically extract and clean product size chart data from multiple sources. The tool then transforms the data into standardized sizes, removes duplicate information and autocorrects missing or incorrect measurements. Amazon also is experimenting with new ways to highlight the most relevant size and measurement details for each customer, such as grouping measurements for their recommended size.
Amazon, like much of the retail industry, has been focusing on AI’s potential in recent months. The company gained a minority ownership position in AI firm Anthropic in September 2023, and AI-generated product recommendations in apparel and health and beauty sectors powered sales during Amazon’s fall Prime Days event.
