From Knowledge Graphs to Multimodal Search for eCommerce sites | CDW21 Masterclass | Volpini&Wang

In this workshop, we showcase a prototype of a multimodal search engine designed for e-commerce websites, utilizing Jina AI and WordLift's product knowledge graph. The session is organized into two parts: first, we create a product knowledge graph from diverse data sources, including PIM systems, merchant feeds, and web crawling; next, we demonstrate how this graph enhances the search experience. The standout insight is that enriching product data with structured information—such as categories, images, and FAQs—greatly boosts SEO and increases organic click-through rates, evidenced by a 35% rise in clicks from a recent implementation. This enhanced data also enables a more powerful internal search, allowing users to perform multimodal queries with both images and text, and filter results by attributes like color and brand.