The Challenge
Customers increasingly rely on third-party validation during their decision-making process, making it a critical aspect of modern shopping. However, the current experience requires them to sift through hundreds of reviews to find the insights they need. Shoppers want quick access to key details about a product's features, fit, and function. This generative AI review summary streamlines the process by consolidating customer feedback, enabling shoppers to make informed purchase decisions more efficiently.
Design Process
The project began with a competitive analysis, a moderated user study, and collaboration with prompt engineers to develop an algorithm that synthesizes customer reviews. The user study revealed that consumers did not perceive bias against generative AI and viewed it as an increasingly ubiquitous feature in shopping experiences. Participants expressed a strong preference for concise, content-rich summaries, particularly favoring quick, scannable attributes outlining a product's features. Collaborating with content writers, machine learning engineers, and product managers, we trained the AI to generate balanced summaries that effectively conveyed both positive and negative customer sentiments. From there, wireframes were designed and deployed for testing on particular clothing categories.
