Nvidia has announced a new blueprint aimed at developing AI-powered retail shopping assistants, designed to enhance customer interactions by integrating the expertise of sales associates, stylists, and designers.
The US tech giant describes its blueprints as essential workflows for both agentic and generative AI applications. Cynthia Countouris, Nvidia's director of product marketing for retail, stated that these blueprints serve as 'starting blocks' for creating custom AI solutions.
The shopping assistant blueprint allows for an end-to-end multimodal capability, enabling the AI to understand both text and images. This technology can handle multiple queries simultaneously, allowing customers to ask about various products in a single interaction.
For instance, a user can inquire whether a retailer stocks tops or shoes while also seeking recommendations for a purse that matches a specific skirt. Countouris emphasized that the assistant can manage ongoing conversations, mimicking interactions with human store associates.
In the home furnishing sector, the assistant can recommend furniture and provide product details, including care instructions. It also allows customers to visualize furniture in their own spaces through Nvidia Omniverse, a real-time AI-based 3D graphics platform.
Countouris highlighted the hyper-personalized experience offered by the assistant, which can incorporate customer engagement context and previous purchase history using retrieval-augmented generation (RAG).
In conjunction with Nvidia's announcement, SoftServe introduced its own genAI shopping assistant, utilizing the new blueprint to enable virtual try-ons for customers. Nvidia has conducted trials of this technology, which has gained traction in enabling virtual product interactions.
Countouris revealed that Nvidia is in discussions with several European retailers regarding the blueprint, with formal announcements anticipated soon. She believes that agentic AI will transform the retail landscape, providing workers with enhanced tools to perform tasks more efficiently.
Looking forward, Countouris predicts that the adoption of agentic AI will expand beyond e-commerce, potentially benefiting supply chain operations and warehouse management.